BRCA_wustl_runs.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of COSMIC genes  3959\n",
      "Num COSMIC genes in this cancer 555\n",
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n",
      "[0, 1, 2, 3, 4, 5, 6, 7, 8]\n"
     ]
    }
   ],
   "source": [
    "import mutex as mex\n",
    "import csv\n",
    "\n",
    "mutationmatrix = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.m2'\n",
    "patientFile = '/Users/jlu96/maf/new/BRCA_wustl/shared_patients.plst'\n",
    "geneFile = None\n",
    "minFreq = 0\n",
    "COSMICFile = '/Users/jlu96/conte/jlu/Analyses/CancerGeneAnalysis/COSMIC/COSMICGenes_OnlyLoss.txt'\n",
    "closer_than_distance = 100000000\n",
    "partition_file = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.ppf9'\n",
    "load_pmm_file = '/Users/jlu96/conte/jlu/Analyses/CancerMutationDistributions/BRCA_wustl-cna-jl-PMM.txt'\n",
    "dna_pmm_comparison_file = '/Users/jlu96/conte/jlu/Analyses/CancerMutationDistributions/BRCA_wustl-cna-jl-PMM-dnacomp.txt'\n",
    "\n",
    "numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "\n",
    "if COSMICFile:\n",
    "    COSMICgenes = set()\n",
    "    with open(COSMICFile, 'rU') as csvfile:\n",
    "        reader = csv.reader(csvfile, delimiter='\\t')\n",
    "        for row in reader:\n",
    "            COSMICgenes.add(*row)\n",
    "    print \"Number of COSMIC genes \", len(COSMICgenes)\n",
    "    genes = (set(genes)).intersection(COSMICgenes)\n",
    "    geneToCases = dict([g for g in geneToCases.items() if g[0] in genes])\n",
    "\n",
    "print \"Num COSMIC genes in this cancer\", len(genes)\n",
    "            \n",
    "\n",
    "\n",
    "cohort_dict, clusterToProp, min_cohort = load_patient_cohorts(partition_file, patientToGenes)\n",
    "\n",
    "print clusterToProp.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number genes is  100\n",
      "number patients is  80\n",
      "Mean is  243.5\n",
      "['FLT3loss', 'BCRloss', 'PPFIBP1loss', 'LHFPloss', 'ERC1loss', 'MLLT1loss', 'SMAD4loss', 'USP6loss', 'PRDM16loss', 'KDSRloss', 'SETD2loss', 'SMARCA4loss', 'ZCCHC8loss', 'GATA2loss', 'HNF1Aloss', 'FOXO1loss', 'CDH11loss', 'CDX2loss', 'SMARCB1loss', 'MALT1loss', 'STK11loss', 'HERPUD1loss', 'MN1loss', 'MYH9loss', 'IRF4loss', 'FHITloss', 'TNFRSF14loss', 'FANCAloss', 'CACNA1Dloss', 'KMT2Closs']\n",
      "Number of genes in cluster 0:  100\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEKCAYAAADticXcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUkP+Pyv1I0jOYDhlAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x142b8dcd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAEKCAYAAAAPVd6lAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZlY4B7mZWeEc5GZmhXOQm5kV7v8DrCM9YH4dMEcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x148562d50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top gene frequencies are  [33, 33, 32, 28, 28, 28, 28, 28, 3, 3]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEKCAYAAADzQPVvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCcUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12702ea50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top gene frequencies are  [96, 96, 96, 95, 95, 95, 93, 86, 43, 38]\n",
      "numbr of genes used is  11\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "# let's look at the smallest cluster\n",
    "c0patients = clusterToProp[0]['Patients']\n",
    "c1patients = clusterToProp[1]['Patients']\n",
    "\n",
    "c0genes, c0geneToCases, c0patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c0patients)\n",
    "c1genes, c1geneToCases, c1patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c1patients)\n",
    "\n",
    "print \"number genes is \", len(c0genes)\n",
    "print \"number patients is \", len(c0patients)\n",
    "print \"Mean is \", clusterToProp[0]['Mean']\n",
    "print list(c0genes)[0:30]\n",
    "print \"Number of genes in cluster 0: \", len(c0genes)\n",
    "\n",
    "\n",
    "\n",
    "pfreq = [len(c0patientToGenes[p]) for p in c0patients]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(pfreq, 100)\n",
    "plt.title(\"Patient Mutation Frequencies in first cluster\")\n",
    "plt.show()\n",
    "\n",
    "gfreq = [len(c0geneToCases[g]) for g in c0geneToCases]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(gfreq, 100)\n",
    "plt.title(\"Gene Mutation Frequencies in first cluster\")\n",
    "plt.show()\n",
    "\n",
    "print \"Top gene frequencies are \", sorted(gfreq, reverse=True)[0:10]\n",
    "\n",
    "gfreq = [len(c1geneToCases[g]) for g in c1geneToCases]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(gfreq, 100)\n",
    "plt.title(\"Gene Mutation Frequencies in second cluster\")\n",
    "plt.show()\n",
    "\n",
    "print \"Top gene frequencies are \", sorted(gfreq, reverse=True)[0:10]\n",
    "# let's limit to the genes with at least \n",
    "\n",
    "test_minFreq = 3\n",
    "test_genes = [c for c in c0genes if len(c0geneToCases[c]) >= test_minFreq]\n",
    "print \"numbr of genes used is \", len(test_genes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of pairs to test:  47\n",
      "Number of pairs is  47\n",
      "number of pairs is  47\n",
      "Original pairs  47 . Min cohort pairs:  47\n",
      "Writing to file...\n"
     ]
    }
   ],
   "source": [
    "import mutex_triangles as met\n",
    "import chisquared as chi\n",
    "import bingenesbypairs as bgbp\n",
    "compute_mutex = True\n",
    "\n",
    "cpairfile = '/Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-cpairs-min_cohort.txt'\n",
    "test_minFreq = 3\n",
    "\n",
    "\n",
    "genepairs = bgbp.getgenepairs(c0geneToCases, test_genes, test_minFreq=test_minFreq, closer_than_distance=closer_than_distance)\n",
    "print \"Number of pairs is \", len(genepairs)\n",
    "\n",
    "cpairsdict, cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, genepairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(cpairsdict)\n",
    "cpairsdict = chi.add_BinomP_min_cohort_all_pairs(cpairsdict, geneToCases, patientToGenes, cohort_dict, cohort_dict[4])\n",
    "\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "\n",
    "met.writeanydict(cpairsdict, cpairfile, fieldnames=fieldnames)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apparently some of the patients in the first cluster have these pairs mutated.\n",
    "HERPUD1loss\tCBFBloss\n",
    "CDH11loss\tMAFloss\n",
    "CBFBloss\tMAFloss\n",
    "FANCAloss\tMAFloss\n",
    "CDH1loss\tHERPUD1loss\n",
    "CDH1loss\tMAFloss\n",
    "HERPUD1loss\tMAFloss\n",
    "CDH11loss\tCYLDloss\n",
    "CDH11loss\tFANCAloss\n",
    "FANCAloss\tCBFBloss\n",
    "\n",
    "HERPUD1loss\t16\t56932048\t56944863\n",
    "CBFBloss\t16\t67029116\t67101058\n",
    "CDH11loss\t16\t64943753\t65126112\n",
    "MAFloss\t16\t79585843\t79600714\n",
    "FANCAloss\t16\t89737549\t89816657\n",
    "CDH1loss\t16\t68737225\t68835548\n",
    "PTENloss 10\t87863113\t87971930\n",
    "CDKN1Bloss\t12\t12715058\t12722371\n",
    "\n",
    "Seems like the 10 MB filter didn't work.\n",
    "\n",
    "\n",
    "Is it possible that those patients just lost a big chunk of their DNA?\n",
    "\n",
    "See cell below: seems like most of the guys with the co-occurrence are just mutated crazily.\n",
    "\n",
    "When we look at the probabilities of the genes minus the highly mutated patients (>400),\n",
    "(i.e. the 47 smallest patients),  only one pair remains:\n",
    "FANCA Aand MAF loss.\n",
    "The others are simply not mutated there. Likely that we have some false positives?\n",
    "\n",
    "\n",
    "Let's look at the gene segments for BRCA_wustl:\n",
    "FANCA/CBFA2T3 are in the same\n",
    " ZNF469_ZFPM1_MIR5189_ZC3H18_IL17C_CYBA_MVD_SNAI3_RNF166_CTU2_PIEZO1_MIR4722_CDT1_APRT_GALNS_TRAPPC2L_PABPN1L_CBFA2T3_ACSF3_LINC00304_CDH15_SLC22A31_ZNF778_ANKRD11_SPG7_RPL13_SNORD68_CPNE7_DPEP1_CHMP1A_SPATA33_CDK10_SPATA2L_VPS9D1_ZNF276_FANCA_SPIRE2_TCF25_MC1R_TUBB3_DEF8_CENPBD1_AFG3L1P_DBNDD1_GAS8_URAHP_PRDM7_TUBB8P7_FAM157Clos\n",
    "\n",
    "\n",
    "CDH11 and CBFB are in the same segment CDH11_LINC00922_RNA5SP428_CDH5_LINC00920_BEAN1_TK2_CKLF_CMTM1_CMTM2_CMTM3_CMTM4_DYNC1LI2_CCDC79_NAE1_CA7_PDP2_CDH16_RRAD_FAM96B_CES2_CES3_CES4A_RN7SL543P_CBFB_C16orf70_B3GNT9_TRADD_FBXL8_HSF4_NOL3_KIAA0895L_EXOC3L1_E2F4_ELMO3_MIR328_LRRC29_TMEM208_FHOD1_SLC9A5_PLEKHG4_KCTD19_RN7SKP118_LRRC36_TPPP3_ZDHHC1_HSD11B2_ATP6V0D1_AGRP_FAM65A_CTCF_RLTPR_ACD_PARD6A_ENKD1_C16orf86_GFOD2_RANBP10_TSNAXIP1_CENPT_THAP11_NUTF2_EDC4_NRN1L_PSKH1_CTRL_PSMB10_LCAT_SLC12A4_DPEP3_DPEP2_DDX28_NFATC3_ESRP2_PLA2G15_SLC7A6_SLC7A6OS_PRMT7_SMPD3_ZFP90_CDH3_CDH1_RNA5SP429loss\n",
    "\n",
    "HERPUD loss:\n",
    "GNAO1_MIR3935_AMFR_NUDT21_OGFOD1_BBS2_MT4_MT3_MT2A_MT1L_MT1E_MT1M_MT1JP_MT1A_MT1DP_MT1B_MT1F_MT1G_MT1H_MT1X_NUP93_SLC12A3_HERPUD1_CETP_NLRC5_CPNE2_FAM192A_RSPRY1_ARL2BP_PLLP_CCL22_CX3CL1_CCL17_CIAPIN1_COQ9_POLR2C_DOK4_CCDC102Alos\n",
    "\n",
    "\n",
    "Let's do a 100MB filter.\n",
    "Then, let's segment.\n",
    "\n",
    "Most genes are mutated only once.\n",
    "\n",
    "Let's confirm that they actually lost a segment of their DNA.\n",
    "\n",
    "There was a segment deletion at chr16:88525832-9035475. Even so, they're not contained in it...\n",
    "\n",
    "They all happened to have deletions at that segment. See 2 cells down."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32\n",
      "[479, 479, 451, 537, 453, 139, 480, 151, 478, 453, 486, 481, 480, 484, 490, 478, 451, 481, 453, 444, 484, 569, 453, 478, 483, 480, 483, 479, 451, 480, 451, 126]\n",
      "{'TCGA-E2-A15L': 9, 'TCGA-EW-A1PG': 9, 'TCGA-A2-A0D3': 9, 'TCGA-E2-A15D': 9, 'TCGA-BH-A0EA': 9, 'TCGA-AR-A2LQ': 1, 'TCGA-AO-A12B': 9, 'TCGA-BH-A28O': 1, 'TCGA-BH-A0H3': 9, 'TCGA-BH-A0DV': 9, 'TCGA-EW-A1IY': 9, 'TCGA-E2-A570': 9, 'TCGA-B6-A0I5': 9, 'TCGA-BH-A2L8': 9, 'TCGA-E9-A227': 9, 'TCGA-E2-A3DX': 9, 'TCGA-D8-A141': 9, 'TCGA-A7-A5ZW': 9, 'TCGA-BH-A0DO': 9, 'TCGA-AC-A3YI': 1, 'TCGA-BH-A0BS': 9, 'TCGA-D8-A27P': 9, 'TCGA-A8-A08C': 9, 'TCGA-D8-A27K': 9, 'TCGA-BH-A0DT': 9, 'TCGA-E2-A1B4': 9, 'TCGA-A2-A0T5': 9, 'TCGA-OL-A5RX': 9, 'TCGA-GM-A2DN': 9, 'TCGA-BH-A0HO': 9, 'TCGA-AO-A125': 9, 'TCGA-AO-A0JC': 1}\n"
     ]
    }
   ],
   "source": [
    "genes1 = [\"CDH11loss\", \"CBFBloss\", \"FANCAloss\", \"CDH1loss\", \"CDH1loss\", \"HERPUD1loss\", \"CDH11loss\", \"CDH11loss\", \"FANCAloss\"]\n",
    "genes2 = [\"MAFloss\", \"MAFloss\", \"MAFloss\", \"HERPUD1loss\", \"MAFloss\", \"MAFloss\", \"CYLDloss\", \"FANCAloss\", \"CBFBloss\"]  \n",
    "\n",
    "genepairs = zip(genes1, genes2)\n",
    "cooccurPatients = {}\n",
    "for patient in c0patients:\n",
    "    for gene1, gene2 in genepairs:\n",
    "        if gene1 in patientToGenes[patient] and gene2 in patientToGenes[patient]:\n",
    "            if patient not in cooccurPatients:\n",
    "                cooccurPatients[patient] = 0\n",
    "            cooccurPatients[patient] += 1\n",
    "            \n",
    "cooccurPatients = dict(sorted(cooccurPatients.items(), key=lambda entry: len(patientToGenes[entry[0]]), reverse=True))\n",
    "print len(cooccurPatients)\n",
    "print [len(patientToGenes[p]) for p in cooccurPatients]\n",
    "print cooccurPatients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[60]\n",
      "['TCGA-E2-A15L', 'TCGA-EW-A1PG', 'TCGA-A2-A0D3', 'TCGA-E2-A15D', 'TCGA-BH-A0EA', 'TCGA-AR-A2LQ', 'TCGA-AO-A12B', 'TCGA-BH-A28O', 'TCGA-BH-A0H3', 'TCGA-BH-A0DV', 'TCGA-EW-A1IY', 'TCGA-E2-A570', 'TCGA-B6-A0I5', 'TCGA-BH-A2L8', 'TCGA-E9-A227', 'TCGA-E2-A3DX', 'TCGA-D8-A141', 'TCGA-A7-A5ZW', 'TCGA-BH-A0DO', 'TCGA-AC-A3YI', 'TCGA-BH-A0BS', 'TCGA-D8-A27P', 'TCGA-A8-A08C', 'TCGA-D8-A27K', 'TCGA-BH-A0DT', 'TCGA-E2-A1B4', 'TCGA-A2-A0T5', 'TCGA-OL-A5RX', 'TCGA-GM-A2DN', 'TCGA-BH-A0HO', 'TCGA-AO-A125', 'TCGA-AO-A0JC']\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "df = pd.read_csv('/Users/jlu96/maf/new/BRCA_wustl/all_lesions.conf_99.txt', sep='\\t')\n",
    "indices = np.where(df['Unique Name'][:] == \"Deletion Peak 32\")[0]\n",
    "print indices\n",
    "\n",
    "keys = list(df.columns.values)\n",
    "patientToKey = {}\n",
    "for key in keys:\n",
    "    if key[:12] in cooccurPatients:\n",
    "        patientToKey[key[:12]] = key\n",
    "\n",
    "delPatients = []\n",
    "for patient in cooccurPatients:\n",
    "    key = patientToKey[patient]\n",
    "    if df[key][60] > 0:\n",
    "        delPatients.append(patient)\n",
    "print delPatients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Smallest cluster size 47\n",
      "Number of pairs to test:  12\n",
      "Number of pairs is  12\n",
      "number of pairs is  12\n",
      "Original pairs  12 . Min cohort pairs:  12\n",
      "Writing to file...\n"
     ]
    }
   ],
   "source": [
    "# let's try to get the probabilities using the least mutated patients.\n",
    "import mutex_triangles as met\n",
    "import chisquared as chi\n",
    "import bingenesbypairs as bgbp\n",
    "compute_mutex = True\n",
    "\n",
    "cpairfile = '/Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-cpairs-smallest-cohort.txt'\n",
    "test_minFreq = 1\n",
    "\n",
    "newc0patients = [p for p in c0patients if len(patientToGenes[p]) < 400]\n",
    "print \"Smallest cluster size\", len(newc0patients)\n",
    "newc0geneToCases = c0geneToCases.copy()\n",
    "for g in newc0geneToCases:\n",
    "    newc0geneToCases[g] = newc0geneToCases[g].intersection(set(newc0patients))\n",
    "\n",
    "genepairs = bgbp.getgenepairs(newc0geneToCases, test_genes, test_minFreq=test_minFreq, closer_than_distance=closer_than_distance)\n",
    "print \"Number of pairs is \", len(genepairs)\n",
    "\n",
    "cpairsdict, cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, genepairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(cpairsdict)\n",
    "cpairsdict = chi.add_BinomP_min_cohort_all_pairs(cpairsdict, geneToCases, patientToGenes, cohort_dict, newc0patients)\n",
    "\n",
    "\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "\n",
    "met.writeanydict(cpairsdict, cpairfile, fieldnames=fieldnames)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, let's take a look at the segment data.\n",
    "\n",
    "BRCA_wustl was segmented. Here are the output from there:\n",
    "cna file size:  60246810\n",
    "time used to read in:  313.754554033\n",
    "Time used to convert genes:  73.9540939331\n",
    "22771 genes were loaded\n",
    "2005 genes were not loaded\n",
    "Time used to get gene positions:  6.44031000137\n",
    "Binning genes.\n",
    "Adjacent concordance threshold:  0.99\n",
    "Distance threshold:  10000000.0\n",
    "Number of position segments  1921\n",
    "Number of segments  1921\n",
    "Number of genes  24776\n",
    "Time used to bin genes:  214.709136009\n",
    "****************\n",
    "WRITING SEGMENT MATRIX:\n",
    "-----------------------\n",
    "Total number of segment alterations:  962262\n",
    "712 segments had equal numbers of gain and losses for  a sample were found, written as gain alterations\n",
    "10109  had some gains and some losses, but were overridden by majority\n",
    "Average majority ratio:  0.778375217298\n",
    "Standard Dev of majority ratios:  0.137182333658\n",
    "*****************\n",
    "********************\n",
    "WRITING SEG2GENES:\n",
    "--------------------\n",
    "Number of genes  22771\n",
    "Number of segments  1921\n",
    "Average concordance:  0.988522560372\n",
    "Standard deviation of concordance:  0.0106270392535\n",
    "Average nonzero concordances:  0.973493456503\n",
    "Standard deviation of nonzero concordances:  0.0250897363908\n",
    "Segments matrix written to:  ./BRCA_wustl-seg-gl.m2\n",
    "Gene to segments written to:  ./BRCA_wustl-seg-gl.gene2seg\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score is  376.0\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnoAAAEnCAYAAAA+Uu64AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/0dO1Fl4JXkUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x178859b90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/axes/_axes.py:475: UserWarning: No labelled objects found. Use label='...' kwarg on individual plots.\n",
      "  warnings.warn(\"No labelled objects found. \"\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEYCAYAAAByXKB5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nM7OO/H/fXVj0k7+jLwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14f169890>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Let's now look at the patient mutation distribution of BRCA, segmented. Is it broad? Does it make sense to partition over this?\n",
    "%matplotlib inline\n",
    "\n",
    "import mutex as mex\n",
    "\n",
    "mutationmatrix = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-seg-jl.m2'\n",
    "numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "mex.graph_mutation_distribution(numCases, genes, geneToCases, filename='Breast Cancer', top_percentile=10, bins=100)\n",
    "mex.graph_patient_distribution(numGenes, patients, patientToGenes, filename='Patient Mutation Distribution', top_percentile=10, bins=100)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial Lambda is  [290.66666666666674, 608.0]\n",
      "Initial p_k is [0.5, 0.5]\n",
      "ll_new is  -42340.9497283 missing is  0\n",
      "ll_new is  -32863.0451459 missing is  0\n",
      "ll_new is  -32066.0198416 missing is  0\n",
      "ll_new is  -31694.8951003 missing is  0\n",
      "ll_new is  -31548.4326697 missing is  0\n",
      "ll_new is  -31484.5649264 missing is  0\n",
      "ll_new is  -31430.4155633 missing is  0\n",
      "ll_new is  -31385.1729503 missing is  0\n",
      "ll_new is  -31375.9516585 missing is  0\n",
      "ll_new is  -31375.2333736 missing is  0\n",
      "ll_new is  -31375.1622617 missing is  0\n",
      "ll_new is  -31375.1516255 missing is  0\n",
      "ll_new is  -31375.149841 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  621.266521169 0.625132291694\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  621.266666158 0.62513197999\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  621.266885775 0.625131507848\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  621.275165315 0.625113708171\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  149.60061249 0.374756621839\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  149.659314728 0.374878308007\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  621.27375492 0.625116740288\n",
      "LL: -31375.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  621.3 0.63\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10c8ae250>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEPCAYAAABRHfM8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwxhjTEQsYBhjjImIBQxjjDER+f8NIsH7ZKFc2AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1276ad710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "Initial Lambda is  [200.0, 444.0, 677.0]\n",
      "Initial p_k is [0.3333333333333333, 0.3333333333333333, 0.3333333333333333]\n",
      "ll_new is  -24631.9003883 missing is  0\n",
      "ll_new is  -18888.2161746 missing is  0\n",
      "ll_new is  -18060.0069886 missing is  0\n",
      "ll_new is  -17682.679785 missing is  0\n",
      "ll_new is  -17520.9105549 missing is  0\n",
      "ll_new is  -17441.2603873 missing is  0\n",
      "ll_new is  -17402.4260438 missing is  0\n",
      "ll_new is  -17385.4008095 missing is  0\n",
      "ll_new is  -17376.5580412 missing is  0\n",
      "ll_new is  -17370.2862189 missing is  0\n",
      "ll_new is  -17363.2775968 missing is  0\n",
      "ll_new is  -17355.3099628 missing is  0\n",
      "ll_new is  -17347.4095318 missing is  0\n",
      "ll_new is  -17340.5785293 missing is  0\n",
      "ll_new is  -17335.6022379 missing is  0\n",
      "ll_new is  -17330.7613079 missing is  0\n",
      "ll_new is  -17322.9743825 missing is  0\n",
      "ll_new is  -17312.7111604 missing is  0\n",
      "ll_new is  -17307.4046622 missing is  0\n",
      "ll_new is  -17305.993743 missing is  0\n",
      "ll_new is  -17305.6529101 missing is  0\n",
      "ll_new is  -17305.5561195 missing is  0\n",
      "ll_new is  -17305.5257626 missing is  0\n",
      "ll_new is  -17305.5157642 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  706.65788033 0.443938563428\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  706.617648133 0.444026472864\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  706.624446429 0.444011617638\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  706.639522563 0.443978675059\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  85.6670688663 0.235029460211\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  344.593308474 0.321019529935\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  344.559877189 0.320953912785\n",
      "LL: -17306.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  706.7 0.44\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmVlNnEDMzKwmTiBmZlaT/w8jIi6UUlhURgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x172823510>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAEPCAYAAACHuClZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZez/A2CAmCZ+JhwpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x16dff5a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "Initial Lambda is  [159.0, 351.60000000000002, 537.19999999999993, 724.0]\n",
      "Initial p_k is [0.25, 0.25, 0.25, 0.25]\n",
      "ll_new is  -17951.4787793 missing is  0\n",
      "ll_new is  -14345.5561838 missing is  0\n",
      "ll_new is  -13623.0088156 missing is  0\n",
      "ll_new is  -13371.6613405 missing is  0\n",
      "ll_new is  -13192.3759718 missing is  0\n",
      "ll_new is  -13031.379367 missing is  0\n",
      "ll_new is  -12935.1614186 missing is  0\n",
      "ll_new is  -12867.0478855 missing is  0\n",
      "ll_new is  -12806.7593869 missing is  0\n",
      "ll_new is  -12761.4839843 missing is  0\n",
      "ll_new is  -12726.0849954 missing is  0\n",
      "ll_new is  -12696.9301195 missing is  0\n",
      "ll_new is  -12671.8340375 missing is  0\n",
      "ll_new is  -12651.4001002 missing is  0\n",
      "ll_new is  -12632.4464327 missing is  0\n",
      "ll_new is  -12605.0149371 missing is  0\n",
      "ll_new is  -12561.8956876 missing is  0\n",
      "ll_new is  -12511.8056523 missing is  0\n",
      "ll_new is  -12471.7239357 missing is  0\n",
      "ll_new is  -12445.2685888 missing is  0\n",
      "ll_new is  -12436.0482587 missing is  0\n",
      "ll_new is  -12432.4238377 missing is  0\n",
      "ll_new is  -12430.4902361 missing is  0\n",
      "ll_new is  -12429.0493514 missing is  0\n",
      "ll_new is  -12427.6147302 missing is  0\n",
      "ll_new is  -12425.9630219 missing is  0\n",
      "ll_new is  -12424.132292 missing is  0\n",
      "ll_new is  -12422.4722782 missing is  0\n",
      "ll_new is  -12421.2803223 missing is  0\n",
      "ll_new is  -12420.5227642 missing is  0\n",
      "ll_new is  -12420.0530735 missing is  0\n",
      "ll_new is  -12419.7613937 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  742.681522454 0.366708328076\n",
      "LL: -12420.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  742.730580814 0.366599288776\n",
      "LL: -12420.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  742.848132884 0.366337881336\n",
      "LL: -12420.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  742.791349355 0.366464177623\n",
      "LL: -12420.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  223.662414475 0.207576894225\n",
      "LL: -12466.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  53.1497083919 0.156764557061\n",
      "LL: -12419.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  53.274623404 0.157084123381\n",
      "LL: -12420.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  742.3 0.37\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n33539ZbZAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x173e6d210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAEPCAYAAACHuClZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nzqIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10da2b050>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "Initial Lambda is  [125.0, 290.66666666666674, 444.0, 608.0, 751.99999999999977]\n",
      "Initial p_k is [0.2, 0.2, 0.2, 0.2, 0.2]\n",
      "ll_new is  -14057.3916834 missing is  0\n",
      "ll_new is  -11950.2036413 missing is  0\n",
      "ll_new is  -11242.5231562 missing is  0\n",
      "ll_new is  -10845.9512687 missing is  0\n",
      "ll_new is  -10651.7580738 missing is  0\n",
      "ll_new is  -10491.5981004 missing is  0\n",
      "ll_new is  -10341.4036119 missing is  0\n",
      "ll_new is  -10251.422193 missing is  0\n",
      "ll_new is  -10230.3532775 missing is  0\n",
      "ll_new is  -10227.0614286 missing is  0\n",
      "ll_new is  -10225.9516822 missing is  0\n",
      "ll_new is  -10225.4781178 missing is  0\n",
      "ll_new is  -10225.249045 missing is  0\n",
      "ll_new is  -10225.1279562 missing is  0\n",
      "ll_new is  -10225.0594294 missing is  0\n",
      "ll_new is  -10225.018534 missing is  0\n",
      "ll_new is  -10224.9930984 missing is  0\n",
      "ll_new is  -10224.9767589 missing is  0\n",
      "ll_new is  -10224.9659928 missing is  0\n",
      "ll_new is  -10224.9587548 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  812.741296284 0.215683168871\n",
      "LL: -10225.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  812.76128083 0.215644996718\n",
      "LL: -10225.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  812.756215215 0.215654672602\n",
      "LL: -10225.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  812.751741624 0.215663217391\n",
      "LL: -10225.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  233.241355681 0.186410590876\n",
      "LL: -11085.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  381.447867183 0.21116611228\n",
      "LL: -10225.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  194.182825906 0.171739011522\n",
      "LL: -10225.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  812.7 0.22\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n58riCcQ551xZPIE455wriycQ55xzZfn//QJZa2bHHR0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x108dbee10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEPCAYAAAC+35gCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1723e0310>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "Initial Lambda is  [102.0, 239.00000000000006, 377.0, 515.00000000000023, 648.0, 771.0]\n",
      "Initial p_k is [0.16666666666666666, 0.16666666666666666, 0.16666666666666666, 0.16666666666666666, 0.16666666666666666, 0.16666666666666666]\n",
      "ll_new is  -11614.0038553 missing is  0\n",
      "ll_new is  -9993.08345582 missing is  0\n",
      "ll_new is  -9595.81475078 missing is  0\n",
      "ll_new is  -9404.11280576 missing is  0\n",
      "ll_new is  -9344.39346006 missing is  0\n",
      "ll_new is  -9322.47226171 missing is  0\n",
      "ll_new is  -9309.3420921 missing is  0\n",
      "ll_new is  -9294.25645742 missing is  0\n",
      "ll_new is  -9275.70250175 missing is  0\n",
      "ll_new is  -9247.5377634 missing is  0\n",
      "ll_new is  -9212.57641006 missing is  0\n",
      "ll_new is  -9180.15689586 missing is  0\n",
      "ll_new is  -9149.37304167 missing is  0\n",
      "ll_new is  -9127.05897359 missing is  0\n",
      "ll_new is  -9112.43905842 missing is  0\n",
      "ll_new is  -9101.1561184 missing is  0\n",
      "ll_new is  -9094.41255174 missing is  0\n",
      "ll_new is  -9089.38342757 missing is  0\n",
      "ll_new is  -9085.25511997 missing is  0\n",
      "ll_new is  -9081.7898884 missing is  0\n",
      "ll_new is  -9078.60614028 missing is  0\n",
      "ll_new is  -9075.36292366 missing is  0\n",
      "ll_new is  -9072.01248111 missing is  0\n",
      "ll_new is  -9068.85008948 missing is  0\n",
      "ll_new is  -9066.19404523 missing is  0\n",
      "ll_new is  -9064.1135269 missing is  0\n",
      "ll_new is  -9062.5065221 missing is  0\n",
      "ll_new is  -9061.24723329 missing is  0\n",
      "ll_new is  -9060.24081329 missing is  0\n",
      "ll_new is  -9059.4243531 missing is  0\n",
      "ll_new is  -9058.75608548 missing is  0\n",
      "ll_new is  -9058.2063941 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  824.784022173 0.191780337428\n",
      "LL: -9058.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  824.69502595 0.191957887365\n",
      "LL: -9058.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  824.790153223 0.191768108548\n",
      "LL: -9058.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  824.548581962 0.192250234508\n",
      "LL: -9058.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  632.218507621 0.206529605957\n",
      "LL: -9055.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  361.076470095 0.18744463059\n",
      "LL: -9508.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  159.109341113 0.155582977115\n",
      "LL: -9135.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  821.7 0.2\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/H9Yc5mNjLoEXgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x113852190>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEPCAYAAAC+35gCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSapm0pAkVTNpSJKqmTQkSdVMGpKkat+OZ5RpovVf0AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14e9f9b10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "Initial Lambda is  [88.0, 200.0, 331.0, 444.0, 567.0, 677.0, 785.0]\n",
      "Initial p_k is [0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285, 0.14285714285714285]\n",
      "ll_new is  -10370.9910406 missing is  0\n",
      "ll_new is  -9250.41494351 missing is  0\n",
      "ll_new is  -9052.9726119 missing is  0\n",
      "ll_new is  -8985.18334303 missing is  0\n",
      "ll_new is  -8880.90012975 missing is  0\n",
      "ll_new is  -8790.12729586 missing is  0\n",
      "ll_new is  -8707.69779787 missing is  0\n",
      "ll_new is  -8628.95372853 missing is  0\n",
      "ll_new is  -8574.69745973 missing is  0\n",
      "ll_new is  -8528.54874532 missing is  0\n",
      "ll_new is  -8483.1582662 missing is  0\n",
      "ll_new is  -8449.20409475 missing is  0\n",
      "ll_new is  -8423.20287939 missing is  0\n",
      "ll_new is  -8399.75710987 missing is  0\n",
      "ll_new is  -8379.9557725 missing is  0\n",
      "ll_new is  -8364.29597414 missing is  0\n",
      "ll_new is  -8349.79142585 missing is  0\n",
      "ll_new is  -8335.54513186 missing is  0\n",
      "ll_new is  -8320.96580914 missing is  0\n",
      "ll_new is  -8305.07460483 missing is  0\n",
      "ll_new is  -8287.13860606 missing is  0\n",
      "ll_new is  -8264.57875945 missing is  0\n",
      "ll_new is  -8234.5211014 missing is  0\n",
      "ll_new is  -8203.13561162 missing is  0\n",
      "ll_new is  -8178.15750036 missing is  0\n",
      "ll_new is  -8166.69104568 missing is  0\n",
      "ll_new is  -8161.59974311 missing is  0\n",
      "ll_new is  -8158.37541353 missing is  0\n",
      "ll_new is  -8155.73036902 missing is  0\n",
      "ll_new is  -8153.25672808 missing is  0\n",
      "ll_new is  -8150.81714596 missing is  0\n",
      "ll_new is  -8148.53326168 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  830.448635213 0.180436305291\n",
      "LL: -8149.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  832.01562705 0.17728909242\n",
      "LL: -8161.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  830.241365367 0.180853792726\n",
      "LL: -8146.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  830.468937678 0.180395500441\n",
      "LL: -8147.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  410.203294004 0.180468369775\n",
      "LL: -8248.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  654.921351661 0.191506871317\n",
      "LL: -8147.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  79.2416171708 0.0906923003146\n",
      "LL: -8315.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  830.2 0.18\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112a65c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEPCAYAAAC+35gCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGKjYAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1252c64d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "Initial Lambda is  [74.777777777777771, 173.55555555555554, 290.66666666666674, 389.11111111111109, 503.0, 608.0, 691.44444444444434, 793.0]\n",
      "Initial p_k is [0.125, 0.125, 0.125, 0.125, 0.125, 0.125, 0.125, 0.125]\n",
      "ll_new is  -9420.72380834 missing is  0\n",
      "ll_new is  -8631.21419257 missing is  0\n",
      "ll_new is  -8454.79315843 missing is  0\n",
      "ll_new is  -8319.43400721 missing is  0\n",
      "ll_new is  -8238.30997097 missing is  0\n",
      "ll_new is  -8189.44507297 missing is  0\n",
      "ll_new is  -8145.42103125 missing is  0\n",
      "ll_new is  -8108.53631228 missing is  0\n",
      "ll_new is  -8083.59986074 missing is  0\n",
      "ll_new is  -8060.80492603 missing is  0\n",
      "ll_new is  -8035.17063546 missing is  0\n",
      "ll_new is  -8007.16788571 missing is  0\n",
      "ll_new is  -7976.13440475 missing is  0\n",
      "ll_new is  -7944.66650172 missing is  0\n",
      "ll_new is  -7914.20595452 missing is  0\n",
      "ll_new is  -7896.87059942 missing is  0\n",
      "ll_new is  -7888.15840659 missing is  0\n",
      "ll_new is  -7881.76053349 missing is  0\n",
      "ll_new is  -7875.95989557 missing is  0\n",
      "ll_new is  -7870.39318233 missing is  0\n",
      "ll_new is  -7865.12511923 missing is  0\n",
      "ll_new is  -7860.63087376 missing is  0\n",
      "ll_new is  -7857.514536 missing is  0\n",
      "ll_new is  -7855.65198815 missing is  0\n",
      "ll_new is  -7854.55974154 missing is  0\n",
      "ll_new is  -7853.89811642 missing is  0\n",
      "ll_new is  -7853.48255122 missing is  0\n",
      "ll_new is  -7853.21405978 missing is  0\n",
      "ll_new is  -7853.03708334 missing is  0\n",
      "ll_new is  -7852.91881361 missing is  0\n",
      "ll_new is  -7852.83903283 missing is  0\n",
      "ll_new is  -7852.78486991 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  906.17708035 0.0646927703411\n",
      "LL: -7853.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  906.880804045 0.0639843642395\n",
      "LL: -7853.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  900.524275446 0.0706769170919\n",
      "LL: -7854.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  906.519786613 0.0643462267965\n",
      "LL: -7853.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  908.171437074 0.0627117173197\n",
      "LL: -7867.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  853.436222213 0.0777851002428\n",
      "LL: -8643.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  832.263393304 0.0920318587225\n",
      "LL: -8534.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  906.2 0.06\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYEgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11dd58bd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEPCAYAAAC+35gCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x142f35ad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "****************************************\n",
      "Begin merging.\n",
      "Possible merged clusters is  [[(7, 6), 0, -608.43994175502064, True, True], [(6, 5), 0, -2172.5256262753528, True, False], [(5, 4), 0, -3641.207795211797, True, False], [(4, 3), 0, -3830.5226231385454, True, False], [(1, 0), 0, -5815.4367252437623, True, False], [(3, 2), 0, -6040.9530485553159, True, False], [(2, 1), 0, -6458.9016825155932, True, False], [(6, 7), 0, -1411.8201256583759, False, False], [(5, 6), 0, -1953.3087742690441, False, False], [(4, 5), 0, -2524.3478254961233, False, False], [(0, 1), 0, -2665.6865930309577, False, False], [(3, 4), 0, -3133.1972741625168, False, False], [(1, 2), 0, -3754.9995235158517, False, False], [(2, 3), 0, -4099.8869686422877, False, False]]\n",
      "Best cluster is  [(7, 6), 0, -608.43994175502064, True, True]\n",
      "MERGING the probs and likelihoods\n",
      "MERGING: HERE ARE OLD VALUES OrderedDict([('Initial Means', array([  74.8,  173.6,  290.7,  389.1,  503. ,  608. ,  691.4,  793. ])), ('Initial p_k', array([ 0.12,  0.12,  0.12,  0.12,  0.12,  0.12,  0.12,  0.12])), ('Initial LL', -9421.0), ('Number', 8), ('OriginalNumber', 8), ('Means', [25.699999999999999, 94.700000000000003, 205.80000000000001, 346.0, 487.60000000000002, 637.5, [773.60000000000002, 906.20000000000005], [906.20000000000005]]), ('Probabilities', [0.080000000000000002, 0.11, 0.13, 0.14999999999999999, 0.14999999999999999, 0.17000000000000001, [0.14999999999999999, 0.059999999999999998], [0.059999999999999998]]), ('Likelihood', -7853.0), ('Classes', [0, 1, 2, 3, 4, 5, 6, 7]), ('AIC', 15738.0), ('BIC', 15815.0), ('Missing', 0), ('MinClusterSize', 52), ('MoreThanMin', 0), ('Merged', False), ('MergeHistory', set([('[906.20000000000005, 773.60000000000002]', '[0.059999999999999998, 0.14999999999999999]', (52, 144), 0, -608.43994175502064, ('Num classes befpre', 8, ('Cluster too small?', True)))])), ('IsFirst', True)])\n",
      "New classe: [0, 1, 2, 3, 4, 5, 6] VS NEW KEYS [0, 1, 2, 3, 4, 5, 6]\n",
      "MERGING: HERE ARE NEW VALUES OrderedDict([('Initial Means', array([  74.8,  173.6,  290.7,  389.1,  503. ,  608. ,  691.4])), ('Initial p_k', array([ 0.12,  0.12,  0.12,  0.12,  0.12,  0.12,  0.12])), ('Initial LL', -9421.0), ('Number', 7), ('OriginalNumber', 8), ('Means', [25.699999999999999, 94.700000000000003, 205.80000000000001, 346.0, 487.60000000000002, 637.5, [773.60000000000002, 906.20000000000005]]), ('Probabilities', [0.080000000000000002, 0.11, 0.13, 0.14999999999999999, 0.14999999999999999, 0.17000000000000001, [0.14999999999999999, 0.059999999999999998]]), ('Likelihood', -7853.0), ('Classes', [0, 1, 2, 3, 4, 5, 6]), ('AIC', 15738.0), ('BIC', 15815.0), ('Missing', 0), ('MinClusterSize', 77), ('MoreThanMin', 0), ('Merged', True), ('MergeHistory', set([('[906.20000000000005, 773.60000000000002]', '[0.059999999999999998, 0.14999999999999999]', (52, 144), 0, -608.43994175502064, ('Num classes befpre', 8, ('Cluster too small?', True)))])), ('IsFirst', True)])\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRx1OwUkfNaQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10df81ad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "End merging.\n",
      "****************************************\n",
      "Initial Lambda is  [67.200000000000003, 159.0, 250.0, 351.60000000000002, 444.0, 537.19999999999993, 635.39999999999998, 724.0, 802.80000000000007]\n",
      "Initial p_k is [0.1111111111111111, 0.1111111111111111, 0.1111111111111111, 0.1111111111111111, 0.1111111111111111, 0.1111111111111111, 0.1111111111111111, 0.1111111111111111, 0.1111111111111111]\n",
      "ll_new is  -8883.6458695 missing is  0\n",
      "ll_new is  -8284.91451932 missing is  0\n",
      "ll_new is  -8077.62355432 missing is  0\n",
      "ll_new is  -7959.60690191 missing is  0\n",
      "ll_new is  -7904.90150475 missing is  0\n",
      "ll_new is  -7864.20235395 missing is  0\n",
      "ll_new is  -7824.52070353 missing is  0\n",
      "ll_new is  -7782.0920864 missing is  0\n",
      "ll_new is  -7739.43439791 missing is  0\n",
      "ll_new is  -7695.70892354 missing is  0\n",
      "ll_new is  -7661.82475125 missing is  0\n",
      "ll_new is  -7643.63751248 missing is  0\n",
      "ll_new is  -7630.85054793 missing is  0\n",
      "ll_new is  -7619.29849424 missing is  0\n",
      "ll_new is  -7608.5352535 missing is  0\n",
      "ll_new is  -7599.54190326 missing is  0\n",
      "ll_new is  -7590.92568408 missing is  0\n",
      "ll_new is  -7579.22547231 missing is  0\n",
      "ll_new is  -7563.01763146 missing is  0\n",
      "ll_new is  -7545.72180638 missing is  0\n",
      "ll_new is  -7530.5616318 missing is  0\n",
      "ll_new is  -7516.19065083 missing is  0\n",
      "ll_new is  -7500.81249757 missing is  0\n",
      "ll_new is  -7485.03663519 missing is  0\n",
      "ll_new is  -7471.08984471 missing is  0\n",
      "ll_new is  -7459.01295789 missing is  0\n",
      "ll_new is  -7449.00142015 missing is  0\n",
      "ll_new is  -7442.20619936 missing is  0\n",
      "ll_new is  -7437.96478804 missing is  0\n",
      "ll_new is  -7435.31343079 missing is  0\n",
      "ll_new is  -7433.52400109 missing is  0\n",
      "ll_new is  -7432.18465598 missing is  0\n",
      "**********NO PATIENTS IN CLUSTER  911.572756723 0.0594528415015\n",
      "LL: -7432.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  912.310781383 0.0587702230183\n",
      "LL: -7440.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  912.307744083 0.0587735669242\n",
      "LL: -7437.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  911.902174814 0.0591472932181\n",
      "LL: -7434.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  757.072408106 0.150382643884\n",
      "LL: -7436.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  26.691582638 0.0823810078584\n",
      "LL: -7804.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  655.500534896 0.150196448493\n",
      "LL: -7611.0 Missing patients:  0\n",
      "**********NO PATIENTS IN CLUSTER  911.6 0.06\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XmcW3XZ///Xu8vQ0g5rS6dQWqBQhIqgCIhsA4j0Fiuy\n21tZ7uJyo+DCfd9KwdIC3j/EW0EQkC/KvhRlkUVRpIWyKmuLFSnQkTJdpgu2dIN2tuv3x+dkepom\nk5xMMklmrufjkUcm55x8znWSTK6cz/ksMjOcc865pPqUOwDnnHPVyROIc865gngCcc45VxBPIM45\n5wriCcQ551xBPIE455wriCcQVxEk7SKpXdIt5Y6lkvnr5CqJJxBXMpI+IukXkv4uaZWkDZIWSfq9\npImSajI8rds6JkmaGn0ZH95d+yyiiujAJalW0v9KmivpQ0krJP1J0lHljs2VXr9yB+B6JkkXA1MA\nAc8D04E1QB1wOPBr4BzggHLFWKUWAh8BVpU7EEnbAs8CewF/B34J1ALHA9MlfdXMbi5jiK7EPIG4\nopN0ITAVaAROMbOXMmxzLPA/3RxaNip3APkys1bgrXLHEZlKSB73A6eZWTt0vP8vA7+Q9JiZLSpf\niK6UvArLFZWkXQhfLM3A5zIlDwAzewz4XB7lzZTUnmXdWVEV1Jlpyz8maZqk+ZLWS1om6RVJV0nq\nF20zH7g4esqTUTnt6fuStKWkSZJmS1oraY2k5yV9KUM89VEZUyQdKOkPUZVOu6SROY6zVtLkWHXf\naknzJN0j6ROx7Ta7BhJ7HTq7jUzb30GS7pO0JKpabJR0g6ThncWZ5gRCVdrFqeQBYGbLgSuBgcDE\nBOW5KuNnIK7Y/oPwuZpmZv/obEMza86zzFz1/R3rJX0MeAFoAx4G3gG2AvYgVJldBLQCVwFfBI4A\nbgXmpxcqaRvgCWA/4BXgJsKPrnHA3ZLGmtnkDPEcDEwCniFU1Q0BWrIFL0nAn6LnPR/93QrsDNQD\nTwOvZjtmYBYhaafbBvgO0A6sj+1vInAj8CHhNVoAjAG+CoyX9CkzW5At3pi66P6fGda9E90fBVyW\nR1muCnkCccV2aHQ/o0z7PxPYAjjezB6Jr5C0NeFLEzO7OqrDPwK41cyezlDWzwnJ4/tm9tNYOVsA\nDwIXSrrPzF5Le95ngW+Y2a/yjPmjhOTxOzM7KX1llMiyiva/SQzRmdafCNVz55vZsmj5GOAGwpf+\nEWbWFHvOUcCfgauBE/OI+z1gGLAb8Ebaut2i+z3zKMdVKa/CcsWWqgJZWNYoYr+4U8xsleU5/LSk\n7YGvAC/Fk0dUzgbgAsKX879nePqsBMkjbrOYo/29X0BZNxJ+/V9jZr+ILT+H8MPxO/HkEe3nCeAR\nwlnIoDz28XvCa3CJpI7vEklDge9FD7ctIHZXJfwMxPU09wDfBh6UdB/hTOg5M2tIWM4BRD+wJE3N\nsL5/dL9XhnUvJtzX68BsYIKkUcBDhNZNL5tZ1qqvbCRdBJwFPGxm301bfXB0Xy/poAxP3wHoSzhz\nSK82S3cxcCxwMjBb0hPAIOALhB8QOxOqz1wP5QnEFVsToZnpiHLs3MxeknQY4VrHycDpAJLeBC4x\ns3vyLGr76P4Asjc1NsIXZrol+UcMZtYeVR9dTIj5imjVGkm3AZPMbF0+ZUmaQLjm8DIwIcMmqePq\nrAVctuNKj3uJpAOAycDnCWc3ywlJ/BrgbWBZPnG76uRVWK7Ynonujy5SeammoZk+qxmvDZjZX81s\nfLT+EMIX6jDChe9840r1s7jSzPpkufU1s0zlJe7kZ2bvm9n5ZjaScMH/q8Bc4FxC/4qcosR5C/Au\nMN7MPsxyXAZsleO4nsnw3ExxLzOz88xsVzPbwsxGmNl3gFHRJhlb4bmewROIK7ZbCC2OTpKUqXqn\nQ5ae6OlWEurZMzWD/WRnTzSzFjP7i5lNIVRrQejkltIW3ffN8PQXCMmr23upm1lD1AHvCGAdoUqo\nU9HF8QcJjQSOM7OlWTb9C+H1LPVxnRHd313i/bgy8gTiisrM3iU0Ka0B/iBp/0zbSfo3QiuhXF6I\n7r+W9vyjyVBFI+nTkgZkKCfV5DReFfSv6H5U2rapvgx3AZ+U9MNMZ0CSRkf9Xrok6tuxW4ZV2xFa\nlGU6k4g/fwjwKDAYODlH8+lrCQn+Kkl7ZCirJjqTySduSRqcYfnphATynJk9mE9Zrjr5NRBXdGZ2\nedSMdArwkqTnCf0o1hKqkg4Hdie/6o1bCPX1kyTtS2guOobQF+MBIL3Z6/eBIyU9Q+jbsRYYG22/\ngtA6KeUJwlnG5ZL2IZztmJn9b7T+XEJ10qXA6ZKeA5YCOxIunn8S+BIZ+pAktB/wgKQXCdVWi4Gh\nhLOlvmy8JpLNpYRms68Ch2VJAFdFrdDejPqB3Ay8LulPhGsV/QlneYcRjnHvPOIeBCyV9GdCs+B2\nQpXhp4B/AKfkUYarZmbmN7+V5Ea4mH4NMIdQ974BWAT8gdDhsH9s210IX0A3Zyhn7+g5qwnjaT1B\n+KI7k1ANdUZs22OIvhyB9wkJ5A1Cn46dM5T9ZUJHvA+i/belre8PfAt4LipvPSFhPE6oFtsutm19\nVMbFCV+nnYD/JbS8aor20Rgd87Fp2272OhGSbFt0a89wawNGppXz0eh586P9vQf8jXC9pT7PuPsR\nOkrOjV7ntYQkdgEwoNyfP7+V/qbog1AWknYGbic0HTTgRjO7Jmo2+VVCiw4IrVDyqe5wzjnXTcqd\nQOqAOjObHdWlvkIYXuJUYI2ZXVm24JxzznWqrNdAzGwJUZt5M1sr6Q3C6TxU0QipzjnXG1VMK6yo\nNcvHgb9Gi86T9Jqkm3KNBeScc677VUQCiaqv7iOMz7OWcCFvV0LrlCbgZ2UMzznnXAZlvQYCIKk/\nYVC2P5rZzzOs3wV4xMz2SVteEVN6OudctTGzolwiKOsZSDQPwk3AP+LJI21SmxMIzUA3U+4mbKW8\nTZkypewx+PH58fXG4+vJx2ZW3N/d5e5IeAhhyOy/SZoVLbuQMCrpfoSmve8A3yhTfM4557Iodyus\nZ8l8FvTH7o7FOedcMhVxEd1trr6+vtwhlJQfX3XrycfXk4+t2Mp+Eb1QkqxaY3fOuXKRhBXpInq5\nr4E451zFCe17ql+pf2R7AnHOuQyqvYajO5JgVSeQN99cBMA222zBsGFDyhyNc871LlV9EX3QoJ3o\n23d72tpyb+ucc664qjqBQM+pq3TOuWpT9QnEOedceVT1NRDnnOsuDQ0LaW4uXfk1NTB69Iic261Y\nsYKzzz6bxx9/nCFDhnD55ZczYcKE0gXWCU8gzjmXh+ZmqK3N/QVfqDVrFua13be+9S0GDBjAsmXL\nmDVrFscddxz77rsve++dzzT2xZV3FZak7STtLWlA2vKJkh6SNE3SQcUP0TnnHMC6det44IEHuOyy\ny9hyyy055JBDOP7447njjjvKEk+SayD/C7xAbKZASecBvwbGA6cBT0rq/jTonHO9wFtvvUW/fv3Y\nfffdO5btu+++vP7662WJJ0kCOQR4wsw+jC37b2ARcDhhHnMB/1W88JxzzqWsXbuWrbbaapNltbW1\nrFmzpizxJLkGshMwI/UgOtPYGbggGlUXSacChxU1QueccwAMHjyY1atXb7Js1apV1NbWliWeJGcg\nA4H1sceHRvfTY8sagNJdZXLOuV5szJgxtLa2Mm/evI5lr732Gh/96EfLEk+SBLIY+Ejs8WeB1cBr\nsWXbAvEqLuecc0UyaNAgTjzxRC6++GI++OADnn32WR555BFOP/30ssSTpArrCeCs6ML5euALwANm\nFh9IZDdgQRHjc865ilBTk39T20LLz8f111/PxIkT2WGHHRgyZAg33HADe+21V8ni6kze84FI2g14\niXCWAbAWONDM5kbrtwaagFvN7JsliDU9HluwwNiwYQMDB65hxx19MEXnXHFEc2aUO4wuyXYMZZkP\nxMz+KemjwMnRoofMrDG2yWjgRuDuYgTmnHOuslX1jIR+BuKcKwU/A8lPQUOZSBoE7AkMMrNnihGI\nc8656pJoNF5JO0t6AHgfeBmYGVt3mKR/SKovaoTOOecqUpKxsIYDfyW0vvo98Bdiw5oQhjkZRhjS\nxDnnXA+X5AxkCiFBfNbMTgAej680s2bgGcKQJ84553q4JAnkc8DDZvZEJ9s0Ajt2LSTnnHPVIEkC\nGQa8lWObFmBw4eE455yrFkkSyErC4Imd2QNYUng4zjnnqkWSZrzPAl+QNNzMmtJXStoDGAfcVazg\nnHOuUjTMb6C5rXRz2tb0rWH0LqM73ebaa6/l1ltv5e9//zsTJkzglltuKVk8+UiSQP4P+CLwlKTv\nEEbnRdJgwnwgVwEG/KzYQTrnXLk1tzVTO7R0w6avWZ57To+ddtqJyZMn89hjj/Hhh+UftzbJUCYv\nSPo6cAPwh9iqVYTmvC3ARDP7e3FDdM45B3DCCScA8PLLL7NwYekGdsxXop7oZnazpGeBc4CDge0J\nCeQvwLVm9mbxQ3TOORdXKcOsJB7KxMzeAr5Xglicc87lQSrKUFZdlmgoE+ecc+VX8WcgkkYWWmja\nMO/OOeeKqFLOQDqrwppPaFWVNFID+hYakHPOucza2tpoaWmhtbWVtrY2NmzYQL9+/ejbtzxfuZ0l\nkNsLLLMyzq2cc66IavrW5NXUtivl53LZZZdx6aWXdjy+8847mTp1KhdffHHJ4upMWSeUkrQzIVHt\nQEg8N5rZNZK2A34DjCKcCZ1qZu+nPdcnlHLOlYRPKJWfcl9EbwG+Z2ZjgU8B35K0F3AB8LiZjQFm\nRI+dc85VkEJnJBwJ7AdsTegHMsvMFiQtx8yWEI2dZWZrJb0B7ESYc+SIaLPbCBNXeRJxzrkKkiiB\nSBoDXA8clbbKJD0JfDPqJ5KYpF2AjxNNTGVmS6NVSwkjATvnnKsgeScQSbsDzwPbAf8kDK64BKgD\nDiUkleckHWxm85IEEY2ndT/wHTNbE2+iZmYmKWNl5JVXTqW1tZX+/ZsZP/5z1NfXJ9mtc871eDNn\nzmTmzJklKTvvi+iS7gVOAr5LGLakPbauL3AuYUDF+83slLwDkPoTpsj9o5n9PFo2F6g3syXRVLpP\nmtlH0p7nF9GdcyXhF9Hzk+Qi+tGEL/lr4skDwMzazOxq4E/RdnlRONW4CfhHKnlEHgbOjP4+E3gw\nQZzOOee6QZIEUgPMyrHN7Gi7fB0CfAU4UtKs6DYO+DFwjKS3CFVjP05QpnPOuW6Q5CL634Ddc2wz\nOtouL2b2LNmT2GfyLcc551z3S3IG8r/AiZI+l2mlpOOAE6LtnHPO9XBJzkCGAH8Efi9pBvAUG5vY\n1hOqmh4Btpd0RvyJZlbosCjOOVcRFjY0QHPpprSlpoYRozuf0ra5uZlzzjmHGTNmsGLFCkaPHs3l\nl1/OuHHjShdXJ5IkkPjku0eT+WL5+OgWZxQ+rpZzzlWG5mZG1JZuStuFa3KPs9Xa2srIkSN5+umn\nGTlyJH/4wx849dRTmTNnDqNGjSpZbNkkSSATC9xHdbeFc865CrHlllsyZcqUjsfHHXccu+66K6++\n+mplJxAzu7WEcTjnnEto6dKlvPXWW4wdO7Ys+y/3YIrOOecK0NLSwpe//GXOOussxowZU5YYPIE4\n51yVaW9v5/TTT2fAgAFce+21ZYsjUQKRdJCkByQ1SNogqS3t1i6prVTBOudcb2dmnH322Sxfvpz7\n77+/bLMRQrLBFE8mTPIk4F3gJaA1w6Z+0dw550rknHPOYe7cuUyfPp0tttiirLEkaYV1CbAO+FzU\ng9w553qPmpq8mtp2pfxc3n33XW688UYGDBhAXV1dx/Ibb7yRCRMmlC62LJIkkNHAbZ48nHO9Ua5O\nft1h1KhRtLe3596wmyS5BrIMKGE3TOecc9UkSQK5lzBCbpLRdp1zzvVQSaqwLgYOA+6V9B0zm1+a\nkIqnoWFhx9A1NTUwevSI8gYUs8m4OnmMgeOcc5UmSU/0dZK+ATwJNEh6H1iVZdvdihRflzQ3Q21t\nSBpr1iwsczRpYuPqlPTCnHPOlUiSZryHAY8BA4A24EMyV4F5M17nnOsFklRh/RjoC5wB3J0+ra1z\nzrneJUkC2Re4x8zuLFUwzjnnqkeSVlhrgX+VKhDnnHPVJUkCeRQ4olSBOOecqy5JqrAuAJ6TdD3w\nP2a2rkQxOedcxWlobKS5hL3Aa/r0YfTIkTm3+8pXvsKMGTNYt24dQ4YM4eyzz+aiiy4qWVydSZJA\nfkOoxvpP4HRJb5G9Ge9RRYjNOecqRnN7O7Wx8aeKbc2SJXltN2nSJH79618zYMAA3nzzTY444gj2\n33//ssyLniSBxKuvBgEfL3Isrko0NjTS3hx+ifWp6cPI0bl/NbniaGhooDnqgFpTU8No74Da66TP\nPtivXz922GGHssSS9zUQM+uT762UAbvya29up662jrrauo5E4rpHc3MztbW11NbWdiQS1/t885vf\nZNCgQYwdO5Yf/vCHfOITnyhLHP5l75xzVeb6669n7dq1TJ8+nR/+8Ie8+OKLZYnDE4hzzlUhSdTX\n13PKKacwbdq0ssSQ5BpIB0kjgJ2AjNNhmdnTXQnKOedcflpaWth+++3Lsu9ECUTSscBVwEcyrDbC\ndLdGGPLEOedcES1fvpwZM2Ywfvx4BgwYwPTp07n33nuZPn16WeJJMpjip4BHgOXAtcC5wFPAW8Ch\nwF7R+leLH6ZzzpVXTZ8+eTe1LbT8XCRxww03cM4552BmjBkzhjvuuIMDDjigZHF1JskZyCRgA3Cg\nmS2SdC7wpJldKqkPMBX4L+DC4ofpnHPllU8nv1IbMmQIM2fOLHcYHZJcRD8YeNjMFqU/PxqZdwrw\nBnBp8cJzzjlXqZKcgWwNvBt73EzoUAiAmZmk54AJRYrNuV4t1WmwnB0GK3lWT1d+Sc5AlgPbpj1O\n/1T3B7bsalDOuY2dBsvZYTA1q2dt7Qi836JLlySBvMWmCeMvwDGS9gSQNBw4EXi7eOE555yrVEkS\nyB+BIyRtFz2+mnC28aqkl4C5wA7Az4sbonPOuUqUJIH8P8KAiq0AZvYccDLwDrAPsBj4TzO7LUkA\nkm6WtFTSnNiyqZIWSpoV3bp/mEnnnHOdyvsiupmtBv6atux3wO+6GMMtwC+A2+NFA1ea2ZVdLNs5\n51yJlH0sLDN7BliZYZW6OxbnnHP561ICkXS8pJ9LulrSScUKKnKepNck3SRpmyKX7Zxzros6rcKS\nNB74H2CymT2Vtu5W4IzYovMkPWhmJxYhrl+ysUPiZcDPgLOLUK5zzhUkPpFaKSSdnO3tt99mn332\n4ZRTTuGOO+4oWVydyXUN5AvA/sAmg81L+jwheawjDK64Fvga8EVJ/25md3clKDNbFtvXrwljbG3m\nyiun0traSv/+zYwf/znq6+tzlt0wv4HmtmhGt741jN4lvw5ahT6vu8Xnbc53juXeprGxgfb28F4u\nWdJEXd1w+vSpYeTIynxPu1u882BjYxNjx3rnQdg4kVqpLFmTbJytb33rWxx44IFIndf2z5w5s2TD\nn+RKIAcCz5jZh2nLJ6buzexeAEm3A/8E/h3oUgKRNNzMmqKHJwBzMm13/vlT2bBhAwMHrmHHHYfk\nVXZzWzO1Q2sBWLN8Td4xFfq87haft7mUA79Vs/b2Zurqwnu5aFEDdXW1LFlSue9pd0t1HgRoaWnK\nsbUrh3vuuYdtt92Wvffem3nz5nW6bX19/SY/ri+55JKixZHrGkgd8PcMy48gXPi+P7XAzJYAfyDh\nXOmSpgHPA3tKWiBpInCFpL9Jei3a1/eSlOmccz3V6tWrmTJlCldddRVmVtZYcp2BbEsY86qDpFHR\n8keiQRTj3iFUe+XNzDKNnXVzkjKcc663mDx5Ml/96lfZcccdc1ZflVquBLIWSK8ATc3ePivLc9Z3\nKSLnnHMZzZ49mxkzZjBrVvj6rfQzkDnAcZJqzSxVSXxCdP9chu13AbzS1DnnSuCpp55i/vz5jIwa\nx6xdu5a2tjbeeOMNXn755W6PJ1cCuZMwhMlTUbPdPYEvA0uAJ+MbKpxLHUpab3XnnHPF8fWvf50J\nE0Ktv5nx05/+lPnz53PDDTeUJZ5cCeRmwgi7x7JxkMQW4Dtm1pq27dGEi+7lmZzXOedKqE9Nn8RN\nbZOWn8vAgQMZOHBgx+PBgwczcOBAtt9++5LF1ZlOE4iZtUV9PiYAnwbeAx4ws9kZNh9KGKH34aJH\n6ZxzZZakk193mTJlSln3n3MwRTNrI1Rl3Zlju2nAtCLFVXQLFoRLM41LljA26s+Rj1QHwsZFjYwd\nOrZU4bleIDXDIFDWWQadK5ayD6bYXVKdo1rSK95yPS/qQNjS1lKawFyvkZphsNyzDDpXLL0mgTjn\nnCsuTyDOOecK4gnEOedcQTyBOOecK4gnEOeccwXJmkAkrZT0/djjKZIO756wnHPOVbrO+oFsDQyI\nPZ4CGPB0SSNyzjlXFTpLIMvYfCTeHmXBogUdfzc1NTF8+HCgfDMOLmxooGMquJoaRpSpo1kxZjVM\nTf+ZdJrOalaqjoILFmz8nDY2NjJ2bNc6tMZnHKypgdGje/S/edHEZ7IshXxnxayvr+eFF16gX7/w\n9T1ixAjeeOONksXVmc4SyF+AMyS1s3GE3fp8xp83s0tzblQB4rMMNjQ2lH/GweZmRtSGGBauKd8M\necWY1TA1/Wcpxw6qNKmOggBrivj+xcttael6h9b4jINr1izscnm9RXwmy1LId1ZMSVx33XVMnDgx\n98Yl1lkC+T4wBvh6bFl9dMulKhKIc85Vo3LPA5KS9SK6mb0NfIyQROqjxbcBR+Vxc845VyKTJk1i\n6NChHHrooTz11FNliyPnaLzAPGBeVHU138xmdkNczjnnMrjiiisYO3YsNTU1TJs2jfHjxzN79mx2\n2223bo8l734gZtbHzC4pZTDOOec6d+CBBzJo0CD69+/PGWecwSGHHMKjjz5allhyDueeiaSdgf2A\nbYBVwKtm5lfjnHOuF0nUE13SLpIeA94FHiJcE3kQeFfSY5J2KXqEzjnnAFi1ahWPPfYY69evp7W1\nlbvuuotnnnmGcePGlSWevM9AJNUBzwI7EhLI04TmvcOBw4BjgOck7W9mvaftpnOuV+jTpybvpraF\nlp9LS0sLkydPZu7cufTt25e99tqLhx56iN13371kcXUmSRXWZELyuAD4WXSBHQBJ/YDvAj+JtvtW\nMYOsZJXS+a8apDoXAlk7GBbaATHVySvfzlilFu/8l+qkWoxOgEmUssNgquze1BGxEj5XQ4YM4cUX\nXyx3GB2SVGEdBzxuZj+JJw8AM2s1s58Cj0fb9R5R578RtbUbE4nLKNW5sK62riORZNsm2/qsZUed\nvErZUziJ+OyDH3zwAbW1tUXpBJgshtBhsLZ2RNE/mqmy/SPfuyVJIHXAyzm2eYVQpeWcc66HS5JA\nVgOjcmyzM6FVlnPOuR4uSQJ5BjhZ0iGZVko6CDiFcKHdOedcD5fkIvr/B3wemCnpN8AThFZYdcCR\nwASgPdrOOedcD5d3AjGzVySdROj78e/RLW4FMNHMcl0ncc451wMk6oluZr+XNAo4HvgEYdKpVcCr\nwINmtq74ITrnnKtEiYcyMbO1wF3RzTnnXC9V0FhYzuVr8YLFHX83NTZRN7aujNFkt3jxxo5/8c6I\n8VnoStVJsdCZDOOdFQudAXHBgqbo+bk7BKa2BWhsbGLs2N7RgdBl5wnElZQ1G3W1IWksallU5miy\nM9s421x8uIr4LHSlGsai0JkMizEDYqpDYD4zE8ZnMmxpacqxdc8TT/SlkORHwD333MMll1zCggUL\nqKur49Zbb+XQQw8tWWzZeAJxzrk8xBN2KeT7I+Dxxx/nggsu4Le//S0HHnggTU1NZZuh0BOIc85V\nkSlTpjBlyhQOPPBAAIYPL9/gH4mGcy8FSTdLWippTmzZdpIel/SWpD9L2qacMTrnXCVoa2vjlVde\nYdmyZeyxxx7svPPOnHfeeaxfv74s8ZQ9gQC3AOmD2V9AGLhxDDAjeuycc73a0qVLaWlp4f777+fZ\nZ59l9uzZzJo1ix/96EdliSfvBCLpSUmXFTsAM3sGWJm2+AuEDotE918s9n6dc67aDBw4EIDzzjuP\nYcOGsf3223P++eeXbUrbJGcgBwF9SxVImmFmtjT6eykwrJv265xzFWvbbbdlxIjKaT6dJIHMI4y2\n260sNC8oTxMD55yrMP/xH//BL37xC5YvX87KlSu56qqrGD9+fFliSdIK61fApZJGmdm7pQooslRS\nnZktkTQcWJZpoyuvnEprayv9+zczfvznqK+vL3FYnWuKdezKNDthfPbCpsZGRhRpdrqGxkaa28ME\nTI1NTYyty6+zXvx5TUuWMDx6XpIy4h0FlzQtoW54eF6ldRqMdwhsamqkrq77ZgbsDl3tVNhdnQSr\neSbDmpqagvvb5Ft+PiZPnsx7773HmDFjGDBgAKeddhoXXXRR1u1nzpzJzJkzixTlppIkkN8T5j1/\nVtJPgBeBJWQ4OzCzxi7G9TBwJnBFdP9gpo3OP38qGzZsYODANey445Au7rIIotkJARZm+qDF1jcV\ncXa65vZ2aqMv/JZF+XfWiz+vYdGigsrYpKNgw6KK7TQY7xC4aFH3zgzYHbraqbC7Ogkm6bhYaQrp\n6V8K/fr147rrruO6667La/v6+vpNflxfcsklxYslwbYNsb+v7mQ7I8G1EknTgCOAIZIWABcDPwZ+\nK+lsYD5waoI4nXPOdYMkCeT2PLdLdL3CzCZkWfWZJOU455zrXknmAzmrhHE455yrMpXQkdA551wV\nKmgsLEl7AXsBg8zsjuKG5JxzrhokOgOR9HFJrwCvA/cBt8bW1Uv6QNIXihuic865SpT3GYikMcCT\nhBZWVwNjgH+LbfI0YUiSkwjNcHu1jj4hGfqDJHp+WhmZ+m4k6bdRKYrRf6Qn9u2I9+dobGxkbJH6\nCpVTqu8H+ERUPU2SM5ApwBbAp8zse8BL8ZVm1g78BTigeOFVsVSfj0InoImen15Gqu9GbV0dH7S1\nUVtXR0uRXEMHAAAdiUlEQVSUUKpJqv9IXW0dbR+0dfzd3pL/saT6dtTV1dLe3jP6dqT6c9TW1tJS\nxL5C5ZTq+1FbO4IeckgukiSBHA08YGavd7LNAmDHroXknHOuGiS5iL4tIUF0RoSzFOec61HiVXGl\nkM/wLoMHD0ZSx+MPP/yQb37zm1xzzTWlC6wTSRLIMmD3HNvsTe4k45xzVSc+3Esp5DO8y9q1azv+\nXrduHXV1dZx6avkG6khShTUDGC/pI5lWSjqAUM31WDECc845l919993HsGHDOPTQQ8sWQ5IE8mOg\nDXha0jnAcABJH5X0TcJgi2uBnxY9Suecc5u47bbbOOOMM8oaQ5KhTOZKOhGYBsSHgfxbdP8+cEI3\nDPXunHO92rvvvsvTTz/NLbfcUtY4EvVEN7M/SdoNOAM4GNgeWEVovnuLma0ofojOOefi7rjjDg47\n7DBGjRpV1jgSD2ViZisJHQk7G9K9qi1YFOvMtaiRsUML78wV7xBYzEmk8rFg8cbOevEJo2r69GH0\nyJHdFkdP1tDQQHPUNKendPwrVLyVUlNTE8OHDwe882Ap3H777Vx44YXlDqOwsbB6uua2ZmqHhsl5\nWhq72POpRJNI5bVrs4wTRq1ZsqRb4+jJ4hM59ZSOf4WKt1JqaGjqlgmqeqPnn3+exYsXc8opp5Q7\nlOQJRNJXgInAfsBWwGpgFqEK687ihuecc5Whpia/prZdKT8ft99+OyeddBKDBg0qWSz5SjIWVn/g\nfuDz0aJ24D1gCHAkcKSkU4GTzKx3/xRzzvU4lTKH+w033FDuEDokacY7iZA8/kpIGAPMrA4YABwF\nvBCtv6DYQTrnnKs8SRLIGYR50Y80s6fMrBXAzFrNbCYhqfwTOLPoUTrnnKs4SRLICOBBM9uQaaWZ\nrQceirZzzjnXwyVJIE1A/xzb9AMW59jGOedcD5AkgdwFnCJp60wrJW0DnBxt55xzrodL0oz3UmAf\n4AVJlwFPAUuBYUA9MBl4MdquW82fv5hVq9YD+Q2J7ArT2NBIe3N7olkDi7lfgCXvLaBu5yFA5lkI\nFy/e2HGzT58aRo5MPhukq1wN8xtobgu9FWv61jB6F39/yylrApHUDlj64uj+jujeYssA9gDWE6a9\n7TYtLerotFTKdtq9XXtzO3W1dSxqWVSW/QIsWvw2dXWh496iRZu3Fjdr7li/ZMma7gvSdYt4J981\ny/39LbfOzkCeLrDM9KTjnHOuB8qaQMysvhvjcM45V2V8LCznnMtD/PpLKeR7TWfhwoWcc845PP/8\n89TU1HDyySfz85//nL59u/XKAeAJxDnn8hK//lIK+V7T+fa3v82QIUNoampi5cqVHHPMMVx//fWc\nd955JYstm0QJRGE29/HAvoQOgxn7hZjZxK6H5pxzLt3rr7/O1VdfTU1NDcOGDWPcuHG8/vrrZYkl\nyWCKowjT1uYz4YEnEOecK4Fjjz2Wu+++myOOOIIVK1bwxz/+kR/96EdliSXJGcg1hORxM3A7ocd5\naymCcs45l9nUqVP5zGc+w1ZbbUVbWxtnnXUWxx9/fFliSZJAjgL+bGZfLVUwLrP5TU2sGTgQgMam\nJsbWdV8nvkqwuOldNqx+H4Bl7zUBe+X3vFinwkydDhPHEZVXjA6KC2IzVXbHTIYLFmyc1KmYMwSW\nqlzwToOZmBnHHnssp5xyCi+88AJr1qxh4sSJ/OAHP+CKK67o9niSDGXSCvytVIG47FqimQVr6+po\naW8vdzjdzqyZumGDqRs2GLP8W8GkOhXW1dXS3t71KWpS5bW3d70lTmomw9ra2m6ZyTA1W2Bt7QiK\nubtSlQsbL1rXDq0taeunavLee+/xyiuvcO6559K/f3+22247zjrrLB599NGyxJMkgTwPfLRUgTjn\nnOvckCFDGD58OL/85S9pa2vj/fff57bbbmPfffctSzxJqrAmA89KmmBm00oVkHPOVaKavjUlHT6l\npm/uOW0l8cADD/Df//3fXH755fTr14+jjz6aq666qmRxdSbvBGJmr0r6DPCopG8ArwCrsmzb7QMq\nOudcKVXKNZiDDjqIZ555ptxhAMma8W4NXA5sBRwe3bIpSgKRNB9YDbQBLWZ2YDHKdc4513VJqrCu\nAg4DphNG422i9M14Dag3sxUl3o9zzrmEkiSQ8cBfgGPNrDtH3FXuTZxzznW3JK2wBgDPdXPyMGC6\npJclfa0b9+uccy6HJGcgs4HdShVIFoeYWZOkocDjkuaaWcfVoyuvnEprayurV/+L4447jYMPri9p\nx6aUBYs2dgJbsbCREXuVthNYOTW9u5iBH0Z/d+NMhIsXLO74e3nTezB6127ZL+TugBhfv3hxI3vu\n2XPf/+4W7zzYuKiRsUPze22902F2M2fOZObMmSUpO+mUtr+XdFj8S7yUzKwpul8u6XfAgUDHvs8/\nfyobNmxg4cK3GD16H2BjxyaAlpamzQstgvionMvmlb4TWDlZs22cDbAbZyKM77e9tXs7T8ZnNcw1\n6+Hbb/fs97+7xf+3Whrzf219psLs6uvrqa+v73h8ySWXFK3sJAlkR8JgijMkTQNeJnsz3tu7Gpik\nLYG+ZrZG0iDgs0Dxjtw55zoRBh93nUmSQG6J/X16dMvECIMtdtUw4HfRm9gPuMvM/lyEcp1zrlPd\ne6m3eiVJIPkO0V6UV97M3gH2K0ZZzjnnii9JT/RbSxiHc865KpOkGa9zzjnXwROIc865giQZC+sd\ncl/fEGBm1t39RZxzznWzJBfRReZhRbYhDLAIYZpbbxjvnHO9QJKL6LtkWydpd8Kc6YOAcV0Pyznn\nXKUryjUQM5sHnATsBEwpRpnOOecqW9EuopvZh4Sh3r9UrDKdc85VrmK3wmoFhhe5TOeccxWoaAkk\nGjH3i8CCXNs655yrfkma8U4hczPefsBI4Hhga2BScUJzzjlXyZI04811cXw1cJmZXdGFeJxzzlWJ\nJAnkqCzL24GVwBtmVuo50p1zzlWIJP1AZpYwjpJrWraQQe8OpmnJIvZjj3KHs4mmBeGyUVNjIyPG\nbj4D2/IlSxg8bx4A7zV1fZKsBYvDbH+NTU2MreueGQadi1uwaD7UhImfCp1BMD4zaLbZC1MzFfos\nhaXRa8bCarUWBg+ppZUKPElqbmZEbS20ZOnE39LC8MGDGT54MLR2Pf5mM2rr6mhp796Z/pxLabYW\naofWUju0tmMq2sRlRLMQ1g6tpaUt8/9OaptC9+E61+kZiKSCEoyZ+TeTc871cLmqsFpJNkGUou37\nFhyRc865qpArgTQmKGsQsH0XYnHOOVdFOk0gnQ2gmCKpP3AecFG06N2uh+Wcc67SdekiuqRTgbnA\nTwnVV98HPlKEuJxzzlW4JP1AOkg6hJA0DiLM/3E1cKmZrSxibM455ypYogQSzftxBXBCtOg+YJKZ\nNRQ7MOecc5UtrwQiaXvCUCbfAPoDfwH+y8z+WsLYnKtIixY1MWjQIACamprYa6/9yhzRphZEHVMb\nG1cwduyIMkeTWz4dAlPbZFufZB/eqbB4cvUD2QL4LnABYaDEBuACM7u/G2JzriK1tLQyePBgAJqb\nK69janNzM7W1tbS0LCt3KHlJdfYDaGnsvENgtvVJ9rFm+ZrCAnWbyXUG8iZhpN0VwPeA63y8K+ec\nc5A7gYyM7gX8F/BfknIWamYjc27knHOuquV7EX3b6Oacc84BuTsS9prBFp1zziXjCcI551xBPIE4\n55wriCcQ55xzBSloKBPXeyxZvox5gxcC0LRseZmjcfloalrBoEFNNDW9x36V1cexQ1PTcga9PTj6\n+z26M8x4p8KmpiaGDx++yd/xZfFOh6nZDdOXp+Ra3xP5GYjrVIsZg4cOZfDQobT6DIZVobVVDB48\ngtbW3E3uy6W1BQYPHs7gwcNpbevefcdnMvyg5YPN/o4vi89kGH9ephkOc63viTyBOOecK4gnEOec\ncwWp2AQiaZykuZLelvSDcsfjnHNuUxWZQCT1Ba4FxgF7AxMk7VXeqLrXK3PmlDuEknp1zqvlDqGk\n5vTw92/OnFfKHULJzJnds9+7YqrIBAIcCMwzs/lm1gLcAxxf5pi6VU9PILPmzCp3CCXlCaR6eQLJ\nX6UmkJ2ABbHHC6NlzjnnKkSl9gOxfDa68sqLMTM+9amDGT16n1LH5JxzLkZmeX1XdytJnwKmmtm4\n6PEkoN3MrohtU3mBO+dcFTCzonQSqtQE0o8wmdXRwGLgRWCCmb1R1sCcc851qMgqLDNrlXQu8BjQ\nF7jJk4dzzlWWijwDcc45V/kqtRVWp6q9k6GknSU9Kel1SX+X9O1o+XaSHpf0lqQ/S9om9pxJ0fHO\nlfTZ8kWfP0l9Jc2S9Ej0uMccn6RtJN0n6Q1J/5B0UA87vknR53OOpLslbVHNxyfpZklLJc2JLUt8\nPJL2j16TtyVd3d3HkUmWY/u/6LP5mqQHJG0dW1e8YzOzqroRqrTmAbsA/YHZwF7ljivhMdQB+0V/\nDyZc79kL+Anw/Wj5D4AfR3/vHR1n/+i45wF9yn0ceRzn+cBdwMPR4x5zfMBtwMTo737A1j3l+KIY\n/wlsET3+DXBmNR8fcBjwcWBObFmS40nV1rwIHBj9/SgwrkKP7ZjUewD8uFTHVo1nIFXfydDMlpjZ\n7OjvtcAbhH4uXyB8MRHdfzH6+3hgmpm1mNl8wpt+YLcGnZCkEcDngF8DqRYfPeL4ol9zh5nZzRCu\n2ZnZKnrI8QGrgRZgy6hBy5aExixVe3xm9gywMm1xkuM5SNJwoNbMXoy2uz32nLLJdGxm9riZpYbP\nfgEYEf1d1GOrxgTSozoZStqF8OvhBWCYmS2NVi0FhkV/70g4zpRqOOargP8B4mPA95Tj2xVYLukW\nSa9K+pWkQfSQ4zOzFcDPgEZC4njfzB6nhxxfTNLjSV++iOo4zomEMwoo8rFVYwLpMVf9JQ0G7ge+\nY2Zr4ussnEd2dqwV+zpI+jywzMxmsfHsYxPVfHyEKqtPANeb2SeAdcAF8Q2q+fgkjQa+S6ji2BEY\nLOkr8W2q+fgyyeN4qpKki4BmM7u7FOVXYwJZBOwce7wzm2bOqiCpPyF53GFmD0aLl0qqi9YPB5ZF\ny9OPeUS0rFJ9GviCpHeAacBRku6g5xzfQmChmb0UPb6PkFCW9JDj+yTwvJn9y8xagQeAg+k5x5eS\n5PO4MFo+Im15xR6npLMI1chfji0u6rFVYwJ5GdhD0i6SaoDTgIfLHFMikgTcBPzDzH4eW/Uw4WIl\n0f2DseVfklQjaVdgD8IFr4pkZhea2c5mtivwJeAJMzudnnN8S4AFksZEiz4DvA48Qg84PmAu8ClJ\nA6PP6meAf9Bzji8l0ecxet9XRy3uBJwee05FkTSOUIV8vJmtj60q7rGVuwVBga0O/o3QcmkeMKnc\n8RQQ/6GEawOzgVnRbRywHTAdeAv4M7BN7DkXRsc7Fzi23MeQ4FiPYGMrrB5zfMC+wEvAa4Rf6Fv3\nsOP7PiEpziFcYO5fzcdHOBNeDDQTrqH+RyHHA+wfvSbzgGvKfVxZjm0i8Dbwbuz75fpSHJt3JHTO\nOVeQaqzCcs45VwE8gTjnnCuIJxDnnHMF8QTinHOuIJ5AnHPOFcQTiHPOuYJ4AnElE3X2bJd0S7lj\nqWbRa/hkueNwLp0nkDRR79v1kn4WW3ajpFWSEr9ekk6R9CdJyyQ1S3ovmmfhTklnFDf67pfnl1vR\nOxvFklO7pDXRuGKZtpOkhti2RxRpv0VLipLmR8O+dKYqO2xJujV6vc7MvXXisqfG3td8brleY5dQ\nRU5pW2aHADXAjNiyo4GnbOPwyHmRdCPwVeAD4A/AO4TBBfcCPg8cThg2udpl+3JbCHwEWFXCfbcC\ng4AJwK8yrD+aMHpuK2Euma5+EVvafbF0Vt5HCJ+halaKBPgkm472DGFk6+MJozykD8Xxfgli6NU8\ngWzuKMKXzdPQMdz6rsAvkhQi6VBC8lgAHGxmi9PW9yMM89FjWRiI760S7+YVYBTwNTInkK8BG4An\nCEPgdJXS7kvOzEr9GlYlM3sKeCq+LBpA8HhgtpldWo64epNeX4UlabCk3VM34LOEMWKGRY9PjTad\nH9tuQB5Ffzq6vz89eUDHJEQz0pdL6iPpPyW9FFXNrJX0YrRssy+tVBWSpB0UprZcEj3nOUmHRdsM\nkvQzSY1R9dzrkk7OUNZWkv5H0hOSFkraEFW9PSTpU2nbniUp9euvPq2qYEq0TdbqHklbSvqBpJcl\nrY6O9R+Srpa0Qx6vb0orcAvwSUkfS9vHEMKkOPcBKzI9ubMquFj1y8jo8VTCTH0AZ6Yd85nRNv0l\nnSvpUUnvRq/3vxSmTh2XVn599BqOBHZJK++W2HYZY5S0taTLJb0p6UNJKxSqS4/OsG3qPZoiaT9J\nf5D0vqR1kmZKOjjDc2olTVaYdnlV9D7Nk3SPpE9kes3yFe2zXWHa4wsVplFdH31Gf6wwWnXRpD6v\nylKVluk11sYqsiMkTYg+q+skLY7+n2qi7T4j6ano9Vkp6Q5J22XZz/6S7o/+r9YrVF9ep2hU4LRt\nU5+/XSWdJ+lvkj7I9nktBz8DgZOBmzMsfzvt8QOxv+uJzlA68V50P6bTrTZ3B6E6ppHwi9qAE4Hr\nCYMwfiXDc7YBniPMJHc3sD1hFNzHJB0C/D9gK+AhQvXcBOA3kj5tZi/Eytkb+BHhV90jhFnORhFm\nbvs3SePN7LFo21nAJcAUYD5wa6ycmWnxbVJ9IWlbQvXDxwjJ+ibCQHC7A2cRhrlfRn6MMOvhBYSz\njfNi684kDAL4K+DrOcrIZ92ThEETv8PmVSSzovvtgZ8T3o/HgOWEOTXGA49K+pqZ3RRt+w7hNfxu\n9PiqWHmzO4tRYf7u5wjVoS8SXrOhhB88f5Z0jpndmOF4PkkYKPF54EbC+3sSMEPSfqmzHUkC/kQY\nxv356O9WwlDg9YTP/6sZyk9qGuFz/Sjh83tcFN8OhEEBi62QOUzOI5y9/o7wGTgW+B4wVNLDwJ3A\n74EbCFXgXyZ8Dj4XL0Rhnpz7o/3cRxjs8JPAOcDxkg61MEtguqsJ09b+Prq15TrIblPukSTLfSP8\n+jsxuv2MUKd6UfT4JGAtYcTOE2O3IXmUuyPhC7id8MU9gfAFqU6eMyHa/mVgy9jyLQkjv7YDE9Ke\n0x7drk9b/pVo+fvR/mti61KjAT+Q9pytgO0yxLUTYW6Af2RY104Yrj3T8ewSrb85bfnd0fLrMjxn\nS2CrPF7fVNlPR48fJ5xlDIht8wYwN/r7zmj7wxPEf2u0fmRs2ahMxxRbXwPsmGH5VoSRTv8VjzFa\nNx/4ZyfHulmMhB8F7cAv05bvHr3n64FRseX1sc/KGWnP+Xr6+wHsEy27P0tM22SLN8trmL7PmdHy\nl9h0FNwtCT/eWgkzBib9fz4ry2furExx5HiNp0bLVwJ7pr3Hfyd8ka8gTG+cWifCyL7twL6x5YOj\n974FOCRtP9+Ptn8sy2u3IP5eVtKt11dhmVmjmT1gZqkzjGbgyujxHMIH+t7UNtHtvawFbix3MXAC\n0ED49XkX4XrAKkl/lPRlbd6qK/WL6wIz+yBW1gfAD6KHX82wu3WEsf/j7iZ8wGsJMx42x8p7lvDr\nZ9+0mFdbmM40/VgWEX45fURhrvOCRdVTpxGGn/7vDPv6wMxWF1D0rwhnYqdE+zkM2JNwdlJMnV77\nMLNmy1xluZpQ1bYtcECXAghVJ18B1gCT0vYzD7iG8CWXqZXfs2aW3nDjZsJnJVNc6zMsw8yKdUH6\nB/Gyos/6XYTq9f2LtI+uusbM3kw9iP6XfkP4LDxsYU7y1Doj/FiBcIadcjzhvf+NmT2XVv7PCP+P\nx0jamc39xMze7fphFF+vTyBpjgJeMrMPo8epi9xPZdm+U2Y2k1CFdTgwmfAlvJZwCnwH8KdUPWrk\nE4R/5JkZinua8Gtkvwzr3jKzdWn7bifM8/y+ZT4tTp+BDABJh0j6raQFUR1te1RPf260SVfngD6A\n8I/3dOx1LoYHCdWGX4sef53wY+DWIu4jL5LGRvXX/4zqrFOv4U+jTXbs4i72BAYCr2X5In8ius/0\nWXk5fYGFxg5LCV9wKa8TqtEmSHpW4drYwUW+NmGZ4mHjDKPbZlhXDplibIruX8mwLvUDIv7/lbpm\n9ETatphZGxurxD+eobyKnZyrV18DkVRPOLWHkEw/BryscLEUQh1mG3BaVCdsZnZJkn1Ev0iejW6p\n/R5DmKTnM4T6z6ujVVsDqWlE08tplfQeMCTDbrI1k23NsW6T91/SCYS62Q8IVUINhLObduBIQkLd\nIkt5+domui/qVKBm1izpduB8hQv+JxN+HeY8WyymaN9PED5PMwiJbTXhNUw1Me3qa7h1dN+UZf2S\n6H6bDOuynTmkmjkD4QeIpKOAiwmv5RXRqjWSbiNM5LZu82KSyXK2mfr8982wrhwy/Q+15rEunmxz\nvWdNadvFLcmwrCL06gRC+EK8OG3ZAWx+Kj8lujfCRc8uMbPHJf2QUL1yJBsTyCpgO0l9o18lHRSa\n/Q4hfBmVymWEKotPxk/Zo/3vRHGaHa+M7rt6JpPJr4DzgXsJX9KZLiJnku3/INMXcC4/BAYA9Wa2\nSUMLSZMICaSrUl9am7XciQxP264g0dnN+YSkPJrw/n+DcDa6DZmryCpVqsXgZu911CCh1LrynlVs\nJ9JeXYVlZpeYWR8z6wNcSfjyHBA93iva7D9T25hZMX8RrY3u43XqrxJ+dWX6oj6c8H4Vo+VLNrsT\nLpSnJ48+hAvvmRjJfim+GD3ncElbFhRltkBC3M8QktM7ZjY9j6etJLQs2oSkvoQqoPR/3lRiz3bM\nuxPOIjO10suWgNs6KS+TucCHwL6SMv1iPTK6L9pnxcwazOxmwjGsI7TMqyapHy4jM6z7ZDfsP/Ve\nHJm+IvpxeBjhs1bK/++i69UJJM2RwF9jF5tTb/TMQgqTNE7SidGHI33dYDY23Yx/0aSaE18uaWBs\n+y2BH0cPb6J03gHGSEr9Gko155xKSKiZfgn9iwxfwNlEVUrTCNcBfhqV30GhX85WyUPv8HVC348T\n89z+BWBUVK0Y90Myf9mkvohGZSnvHWB7SfvEF0o6m9DHKJN/ATsov/5FmFkL4ULtVoSzxvh+RgPf\nJlz/uSOf8jJR6MOzW4ZV2xHO7op5/ao7pFox/nva/9Z2wE+6Yf8PElpsTZB0UNq67xJaFU43s4Xp\nT6xkvb0KC+g4hd2PTaun6oEmK7wX8J6Edv0rJT1DmKi+lXBh7ThCXedfgWtTTzCzaZKOJ7Tlf13S\nQ4Qv7S8SPmD3mNm0AuPJJL1F0VWEtuyzJD1A1OSQkDweIbQmSzcd+FLUHn5W9Jyn4i1TMjgX+Cjw\nn4ROiH8mfOHtSviSHU/ufjYZRWchb+bccKOfEho1PCTpN4QE8WnC6z2TjdfIUuWvlfRX4DBJdxKa\nnLYBD5nZHEIfkGOBZyX9llDl+EnC63gf4XpCuunRNn+KPisbCD2pf99J3BcQfrWeK+mAKNYhhM/O\nIODcLrbc2Q94QNKLhDOexYR+JscTzpau6OS5+erO3vxLJN0FnA7MlvQoIQH/G6GRTKYGB8Xc/zpJ\nEwnVq09JupfQPHd/4BjCNZBvlDKGUvAEEqSqFmbGlh1OgWcfkTsJXx7HEJrLHk5oC76ScJp6L/Dr\nDBfMJxA+0BMJv6aN0J/h/8zslwljyNVpapP1ZnajpA2EX0RnEC6mP0PokHcyYfyudN+Jyjma0OhA\nhEScNYGY2fuSPh3t5zRCy6k2QufJmwjHW2ybHW8UyxOSvki4FvYlQtXi44TmwJdmeg7hS+gqYBzh\n/YIQ+xwze0zSeMIZzGmEHw0vEhLRaELfonQ/IlxTGE9INH0IjSyyJhAzW6nQe3wS4Wzre4T366+E\nz0o+1XebFJn2+CXgcsL/xrGEFlHLouXX2MYOpfmUm+k1zLY817p89pfN1witzSYA3yQ0nb2a8CPi\ntFLHaGYPRx17LyS8plsTEscvgcvMLP1ieVdeh26h0EjIOeecS8avgTjnnCuIJxDnnHMF8QTinHOu\nIJ5AnHPOFcQTiHPOuYJ4AnHOOVcQTyDOOecK4gnEOedcQTyBOOecK4gnEOeccwX5/wGixasKlNkD\n+wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x140c74950>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEPCAYAAAC+35gCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1231a3990>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No missing patients in clusters\n",
      "****************************************\n",
      "Begin merging.\n",
      "Possible merged clusters is  [[(8, 7), 0, -559.28530332253808, True, True], [(0, 1), 0, -1859.5771198775881, False, True], [(7, 6), 0, -2223.2661162945515, True, False], [(5, 4), 0, -2602.8421645872531, True, False], [(3, 2), 0, -2815.4089988092196, True, False], [(6, 5), 0, -2977.5941015797398, True, False], [(4, 3), 0, -3009.5854831597771, True, False], [(2, 1), 0, -3413.0386791156843, True, False], [(1, 0), 0, -3845.9732023191755, True, False], [(7, 8), 0, -1473.8042449165134, False, False], [(3, 4), 0, -1801.1157180131947, False, False], [(5, 6), 0, -1820.9029109254566, False, False], [(6, 7), 0, -1829.1573881128322, False, False], [(4, 5), 0, -1988.0026597600256, False, False], [(2, 3), 0, -2020.8144952905745, False, False], [(1, 2), 0, -2220.0416591840108, False, False]]\n",
      "Best cluster is  [(8, 7), 0, -559.28530332253808, True, True]\n",
      "MERGING the probs and likelihoods\n",
      "MERGING: HERE ARE OLD VALUES OrderedDict([('Initial Means', array([  67.2,  159. ,  250. ,  351.6,  444. ,  537.2,  635.4,  724. ,\n",
      "        802.8])), ('Initial p_k', array([ 0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11])), ('Initial LL', -8884.0), ('Number', 9), ('OriginalNumber', 9), ('Means', [22.5, 78.900000000000006, 163.30000000000001, 263.89999999999998, 377.30000000000001, 503.10000000000002, 642.5, [778.0, 911.60000000000002], [911.60000000000002]]), ('Probabilities', [0.070000000000000007, 0.089999999999999997, 0.10000000000000001, 0.10000000000000001, 0.13, 0.13, 0.16, [0.14999999999999999, 0.059999999999999998], [0.059999999999999998]]), ('Likelihood', -7432.0), ('Classes', [0, 1, 2, 3, 4, 5, 6, 7, 8]), ('AIC', 14900.0), ('BIC', 14988.0), ('Missing', 0), ('MinClusterSize', 45), ('MoreThanMin', 0), ('Merged', False), ('MergeHistory', set([('[911.60000000000002, 778.0]', '[0.059999999999999998, 0.14999999999999999]', (45, 149), 0, -559.28530332253808, ('Num classes befpre', 9, ('Cluster too small?', True)))])), ('IsFirst', True)])\n",
      "New classe: [0, 1, 2, 3, 4, 5, 6, 7] VS NEW KEYS [0, 1, 2, 3, 4, 5, 6, 7]\n",
      "MERGING: HERE ARE NEW VALUES OrderedDict([('Initial Means', array([  67.2,  159. ,  250. ,  351.6,  444. ,  537.2,  635.4,  724. ])), ('Initial p_k', array([ 0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11])), ('Initial LL', -8884.0), ('Number', 8), ('OriginalNumber', 9), ('Means', [22.5, 78.900000000000006, 163.30000000000001, 263.89999999999998, 377.30000000000001, 503.10000000000002, 642.5, [778.0, 911.60000000000002]]), ('Probabilities', [0.070000000000000007, 0.089999999999999997, 0.10000000000000001, 0.10000000000000001, 0.13, 0.13, 0.16, [0.14999999999999999, 0.059999999999999998]]), ('Likelihood', -7432.0), ('Classes', [0, 1, 2, 3, 4, 5, 6, 7]), ('AIC', 14900.0), ('BIC', 14988.0), ('Missing', 0), ('MinClusterSize', 68), ('MoreThanMin', 0), ('Merged', True), ('MergeHistory', set([('[911.60000000000002, 778.0]', '[0.059999999999999998, 0.14999999999999999]', (45, 149), 0, -559.28530332253808, ('Num classes befpre', 9, ('Cluster too small?', True)))])), ('IsFirst', True)])\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1467e4b50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "End merging.\n",
      "****************************************\n",
      "****************************************\n",
      "Begin merging.\n",
      "Possible merged clusters is  [[(0, 1), 0, -1859.5771198775881, False, True], [(5, 4), 0, -2602.8421645872531, True, False], [(3, 2), 0, -2815.4089988092196, True, False], [(6, 5), 0, -2977.5941015797398, True, False], [(4, 3), 0, -3009.5854831597771, True, False], [(2, 1), 0, -3413.0386791156843, True, False], [(1, 0), 0, -3845.9732023191755, True, False], [(7, 6), 0, -4679.0246543041067, True, False], [(3, 4), 0, -1801.1157180131947, False, False], [(5, 6), 0, -1820.9029109254566, False, False], [(6, 7), 0, -1829.1573881128322, False, False], [(4, 5), 0, -1988.0026597600256, False, False], [(2, 3), 0, -2020.8144952905745, False, False], [(1, 2), 0, -2220.0416591840108, False, False]]\n",
      "Best cluster is  [(0, 1), 0, -1859.5771198775881, False, True]\n",
      "MERGING the probs and likelihoods\n",
      "MERGING: HERE ARE OLD VALUES OrderedDict([('Initial Means', array([  67.2,  159. ,  250. ,  351.6,  444. ,  537.2,  635.4,  724. ])), ('Initial p_k', array([ 0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11])), ('Initial LL', -8884.0), ('Number', 8), ('OriginalNumber', 9), ('Means', [[22.5], [78.900000000000006, 22.5], 163.30000000000001, 263.89999999999998, 377.30000000000001, 503.10000000000002, 642.5, [778.0, 911.60000000000002]]), ('Probabilities', [[0.070000000000000007], [0.089999999999999997, 0.070000000000000007], 0.10000000000000001, 0.10000000000000001, 0.13, 0.13, 0.16, [0.14999999999999999, 0.059999999999999998]]), ('Likelihood', -7432.0), ('Classes', [0, 1, 2, 3, 4, 5, 6, 7]), ('AIC', 14900.0), ('BIC', 14988.0), ('Missing', 0), ('MinClusterSize', 68), ('MoreThanMin', 0), ('Merged', True), ('MergeHistory', set([('[22.5, 78.900000000000006]', '[0.070000000000000007, 0.089999999999999997]', (68, 86), 0, -1859.5771198775881, ('Num classes befpre', 8, ('Cluster too small?', True))), ('[911.60000000000002, 778.0]', '[0.059999999999999998, 0.14999999999999999]', (45, 149), 0, -559.28530332253808, ('Num classes befpre', 9, ('Cluster too small?', True)))])), ('IsFirst', True)])\n",
      "New classe: [1, 2, 3, 4, 5, 6, 7] VS NEW KEYS [1, 2, 3, 4, 5, 6, 7]\n",
      "MERGING: HERE ARE NEW VALUES OrderedDict([('Initial Means', array([ 159. ,  250. ,  351.6,  444. ,  537.2,  635.4,  724. ])), ('Initial p_k', array([ 0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11])), ('Initial LL', -8884.0), ('Number', 7), ('OriginalNumber', 9), ('Means', [[78.900000000000006, 22.5], 163.30000000000001, 263.89999999999998, 377.30000000000001, 503.10000000000002, 642.5, [778.0, 911.60000000000002]]), ('Probabilities', [[0.089999999999999997, 0.070000000000000007], 0.10000000000000001, 0.10000000000000001, 0.13, 0.13, 0.16, [0.14999999999999999, 0.059999999999999998]]), ('Likelihood', -7432.0), ('Classes', [1, 2, 3, 4, 5, 6, 7]), ('AIC', 14900.0), ('BIC', 14988.0), ('Missing', 0), ('MinClusterSize', 95), ('MoreThanMin', 0), ('Merged', True), ('MergeHistory', set([('[22.5, 78.900000000000006]', '[0.070000000000000007, 0.089999999999999997]', (68, 86), 0, -1859.5771198775881, ('Num classes befpre', 8, ('Cluster too small?', True))), ('[911.60000000000002, 778.0]', '[0.059999999999999998, 0.14999999999999999]', (45, 149), 0, -559.28530332253808, ('Num classes befpre', 9, ('Cluster too small?', True)))])), ('IsFirst', True)])\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEnCAYAAACHcBUBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n7wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x122a7fd90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "End merging.\n",
      "****************************************\n",
      "(OrderedDict([('Initial Means', array([  54.9,  215.8,  313.8,  431.2,  573.8,  684. ,  755.1])), ('Initial p_k', array([ 0.2 ,  0.09,  0.14,  0.24,  0.08,  0.12,  0.12])), ('Initial LL', -10063.0), ('Number', 7), ('OriginalNumber', 7), ('Means', [26.199999999999999, 96.599999999999994, 209.09999999999999, 350.5, 494.10000000000002, 654.89999999999998, 830.20000000000005]), ('Probabilities', [0.080000000000000002, 0.11, 0.13, 0.14999999999999999, 0.14999999999999999, 0.19, 0.17999999999999999]), ('Likelihood', -8146.0), ('Classes', [0, 1, 2, 3, 4, 5, 6]), ('AIC', 16319.0), ('BIC', 16387.0), ('Missing', 0), ('MinClusterSize', 78), ('MoreThanMin', 1), ('Merged', False), ('MergeHistory', set([])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([  90.9,  528. ,  863. ,  236.1,   31.8,  843.2])), ('Initial p_k', array([ 0.11,  0.11,  0.12,  0.21,  0.41,  0.05])), ('Initial LL', -12420.0), ('Number', 6), ('OriginalNumber', 6), ('Means', [38.399999999999999, 139.90000000000001, 282.30000000000001, 445.60000000000002, 632.20000000000005, 821.70000000000005]), ('Probabilities', [0.12, 0.14000000000000001, 0.14999999999999999, 0.19, 0.20999999999999999, 0.20000000000000001]), ('Likelihood', -9055.0), ('Classes', [0, 1, 2, 3, 4, 5]), ('AIC', 18133.0), ('BIC', 18192.0), ('Missing', 0), ('MinClusterSize', 108), ('MoreThanMin', 1), ('Merged', False), ('MergeHistory', set([])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([ 125. ,  290.7,  444. ,  608. ,  752. ])), ('Initial p_k', array([ 0.2,  0.2,  0.2,  0.2,  0.2])), ('Initial LL', -14057.0), ('Number', 5), ('OriginalNumber', 5), ('Means', [52.700000000000003, 194.19999999999999, 381.39999999999998, 598.29999999999995, 812.70000000000005]), ('Probabilities', [0.16, 0.17000000000000001, 0.20999999999999999, 0.25, 0.22]), ('Likelihood', -10225.0), ('Classes', [0, 1, 2, 3, 4]), ('AIC', 20470.0), ('BIC', 20519.0), ('Missing', 0), ('MinClusterSize', 149), ('MoreThanMin', 1), ('Merged', False), ('MergeHistory', set([])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([ 827. ,  495.4,  201.2,  181.2])), ('Initial p_k', array([ 0.24,  0.14,  0.01,  0.61])), ('Initial LL', -22870.0), ('Number', 4), ('OriginalNumber', 4), ('Means', [53.100000000000001, 207.80000000000001, 438.19999999999999, 742.29999999999995]), ('Probabilities', [0.16, 0.20000000000000001, 0.28000000000000003, 0.37]), ('Likelihood', -12419.0), ('Classes', [0, 1, 2, 3]), ('AIC', 24855.0), ('BIC', 24894.0), ('Missing', 0), ('MinClusterSize', 149), ('MoreThanMin', 1), ('Merged', False), ('MergeHistory', set([])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([ 200.,  444.,  677.])), ('Initial p_k', array([ 0.33,  0.33,  0.33])), ('Initial LL', -24632.0), ('Number', 3), ('OriginalNumber', 3), ('Means', [85.700000000000003, 344.60000000000002, 706.70000000000005]), ('Probabilities', [0.23999999999999999, 0.32000000000000001, 0.44]), ('Likelihood', -17306.0), ('Classes', [0, 1, 2]), ('AIC', 34623.0), ('BIC', 34652.0), ('Missing', 0), ('MinClusterSize', 224), ('MoreThanMin', 1), ('Merged', False), ('MergeHistory', set([])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([ 290.7,  608. ])), ('Initial p_k', array([ 0.5,  0.5])), ('Initial LL', -42341.0), ('Number', 2), ('OriginalNumber', 2), ('Means', [149.69999999999999, 621.29999999999995]), ('Probabilities', [0.37, 0.63]), ('Likelihood', -31375.0), ('Classes', [0, 1]), ('AIC', 62758.0), ('BIC', 62778.0), ('Missing', 0), ('MinClusterSize', 358), ('MoreThanMin', 1), ('Merged', False), ('MergeHistory', set([])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([ 159. ,  250. ,  351.6,  444. ,  537.2,  635.4,  724. ])), ('Initial p_k', array([ 0.11,  0.11,  0.11,  0.11,  0.11,  0.11,  0.11])), ('Initial LL', -8884.0), ('Number', 7), ('OriginalNumber', 9), ('Means', [[78.900000000000006, 22.5], 163.30000000000001, 263.89999999999998, 377.30000000000001, 503.10000000000002, 642.5, [778.0, 911.60000000000002]]), ('Probabilities', [[0.089999999999999997, 0.070000000000000007], 0.10000000000000001, 0.10000000000000001, 0.13, 0.13, 0.16, [0.14999999999999999, 0.059999999999999998]]), ('Likelihood', -7432.0), ('Classes', [1, 2, 3, 4, 5, 6, 7]), ('AIC', 14900.0), ('BIC', 14988.0), ('Missing', 0), ('MinClusterSize', 95), ('MoreThanMin', 0), ('Merged', True), ('MergeHistory', set([('[22.5, 78.900000000000006]', '[0.070000000000000007, 0.089999999999999997]', (68, 86), 0, -1859.5771198775881, ('Num classes befpre', 8, ('Cluster too small?', True))), ('[911.60000000000002, 778.0]', '[0.059999999999999998, 0.14999999999999999]', (45, 149), 0, -559.28530332253808, ('Num classes befpre', 9, ('Cluster too small?', True)))])), ('IsFirst', True)]), OrderedDict([('Initial Means', array([  74.8,  173.6,  290.7,  389.1,  503. ,  608. ,  691.4])), ('Initial p_k', array([ 0.12,  0.12,  0.12,  0.12,  0.12,  0.12,  0.12])), ('Initial LL', -9421.0), ('Number', 7), ('OriginalNumber', 8), ('Means', [25.699999999999999, 94.700000000000003, 205.80000000000001, 346.0, 487.60000000000002, 637.5, [773.60000000000002, 906.20000000000005]]), ('Probabilities', [0.080000000000000002, 0.11, 0.13, 0.14999999999999999, 0.14999999999999999, 0.17000000000000001, [0.14999999999999999, 0.059999999999999998]]), ('Likelihood', -7853.0), ('Classes', [0, 1, 2, 3, 4, 5, 6]), ('AIC', 15738.0), ('BIC', 15815.0), ('Missing', 0), ('MinClusterSize', 77), ('MoreThanMin', 0), ('Merged', True), ('MergeHistory', set([('[906.20000000000005, 773.60000000000002]', '[0.059999999999999998, 0.14999999999999999]', (52, 144), 0, -608.43994175502064, ('Num classes befpre', 8, ('Cluster too small?', True)))])), ('IsFirst', True)]))\n"
     ]
    }
   ],
   "source": [
    "# A partition over this space looks good! Let's partition it.\n",
    "\n",
    "run_partitions(mutationmatrix='/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-seg-jl.m2', min_cluster_size=70,\n",
    "              do_cross_val=False, do_pmm_dna=False, maxComp=10)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "[0, 1, 2, 3, 4, 5, 6] [0, 1, 2, 3, 4, 5, 6]\n"
     ]
    }
   ],
   "source": [
    "# when we're ready, let's look at the cooccurring pairs here. EDIT PARTITION FILE\n",
    "\n",
    "import mutex as mex\n",
    "import csv\n",
    "\n",
    "mutationmatrix = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-seg-jl.m2'\n",
    "patientFile = '/Users/jlu96/maf/new/BRCA_wustl/shared_patients.plst'\n",
    "geneFile = None\n",
    "minFreq = 0\n",
    "COSMICFile = '/Users/jlu96/conte/jlu/Analyses/CancerGeneAnalysis/COSMIC/COSMICGenes_OnlyLoss.txt'\n",
    "closer_than_distance = 100000000\n",
    "partition_file = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-seg-jl.ppf9'\n",
    "\n",
    "numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "\n",
    "cohort_dict, clusterToProp, min_cohort = load_patient_cohorts(partition_file, patientToGenes)\n",
    "\n",
    "print clusterToProp.keys(), cohort_dict.keys()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number genes is  1679\n",
      "number patients is  95\n",
      "Mean is  163.3\n",
      "['GAB2loss', 'YPEL2_MIR4729loss', 'KCNK9loss', 'SH3GL2loss', 'KAT6Aloss', 'GLP1R_SAYSD1_KCNK5_KCNK17_KCNK16_KIF6loss', 'LINC00457loss', 'FAN1_MTMR10_TRPM1_MIR211_KLF13loss', 'GOSR1_TBC1D29loss', 'CHMP6_BAIAP2_AATK_MIR657_MIR1250_ENTHD2_C17orf89_SLC38A10_LINC00482_TMEM105_MIR4740_MIR3186_ACTG1_FSCN2_C17orf70_NPLOC4_TSPAN10_PDE6G_OXLD1_CCDC137_ARL16_HGS_MRPL12_SLC25A10_GCGR_FAM195B_PPP1R27_P4HB_ARHGDIA_ALYREF_ANAPC11_NPB_PCYT2_SIRT7_MAFG_PYCR1_MYADML2_NOTUM_ASPSCR1_STRA13_LRRC45_RAC3_DCXR_RFNG_GPS1_DUS1L_FASN_CCDC57_SLC16A3_CSNK1D_CD7_SECTM1_TEX19_UTS2R_OGFOD3_HEXDC_C17orf62_NARFloss', 'RHOT1loss', 'OPRM1_IPCEF1_CNKSR3loss', 'RN7SKP106loss', 'TRIM32_SNORA70C_RN7SKP128_RN7SKP125loss', 'FAR2_ERGIC2_OVCH1_TMTC1loss', 'CNEP1R1_HEATR3_RNY4P3_PAPD5_ADCY7_BRD7_NKD1_SNX20_NOD2_CYLD_RNA5SP426loss', 'PTGFRN_RNA5SP55_CD101_TTF2_MIR942_TRIM45_VTCN1loss', 'SUCLG2loss', 'B9D1_MIR1180_MAPK7_MFAP4_RNF112_SLC47A1_SNORA59B_ALDH3A2_SLC47A2_ALDH3A1_ULK2_AKAP10loss', 'CPPED1_MIR4718_SHISA9loss', 'RNA5SP356loss', 'NDUFA4_PHF14loss', 'GPAM_TECTB_GUCY2GP_ACSL5_ZDHHC6_VTI1A_MIR4295_TCF7L2_HABP2_NRAP_CASP7_PLEKHS1_DCLRE1A_NHLRC2_ADRB1_MIR2110_TDRD1_VWA2_AFAP1L2_RN7SL384P_ABLIM1_FAM160B1_TRUB1loss', 'ASH2L_STAR_LSM1_BAG4loss', 'MTAPloss', 'LEPROTL1_MBOAT4_DCTN6_TUBBP1_RBPMSloss', 'TPCN2_MIR3164loss', 'NFIC_DOHH_FZR1_MFSD12_C19orf71_HMG20B_GIPC3_TBXA2R_CACTIN_PIP5K1C_TJP3_APBA3_MRPL54_RAX2_MATK_ZFR2_ATCAY_RN7SL202P_NMRK2_DAPK3_MIR637_EEF2_PIAS4_ZBTB7A_MAP2K2_CREB3L3_SIRT6_ANKRD24_EBI3_CCDC94_SHD_TMIGD2_FSD1_STAP2_MPND_SH3GL1_CHAF1A_UBXN6_MIR4746_HDGFRP2_PLIN4_PLIN5_LRG1_SEMA6B_TNFAIP8L1loss', 'PRSS23_FZD4loss', 'RNU4ATAC3P_MIR548ASloss']\n",
      "Number of genes in cluster 0:  1679\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAEKCAYAAADdBdT9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10c2bc390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEKCAYAAAAYd05sAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFcoBb2ZWqP8PjL2PXhtVqccAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10abd50d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top gene frequencies are  [67, 67, 66, 66, 66, 66, 65, 65, 65, 65]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEKCAYAAAAYd05sAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nzCrq/wPGJ7Bu29kgYQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10a8386d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of patients in first cluster  102\n",
      "Top gene frequencies are  [67, 67, 67, 67, 66, 66, 66, 65, 65, 65]\n",
      "numbr of genes used is  1679\n"
     ]
    }
   ],
   "source": [
    "c0patients = cohort_dict[0]\n",
    "c1patients = cohort_dict[1]\n",
    "\n",
    "c0genes, c0geneToCases, c0patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c0patients)\n",
    "c1genes, c1geneToCases, c1patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c1patients)\n",
    "\n",
    "print \"number genes is \", len(c0genes)\n",
    "print \"number patients is \", len(c0patients)\n",
    "print \"Mean is \", clusterToProp[0]['Mean']\n",
    "print list(c0genes)[0:30]\n",
    "print \"Number of genes in cluster 0: \", len(c0genes)\n",
    "\n",
    "\n",
    "\n",
    "pfreq = [len(c0patientToGenes[p]) for p in c0patients]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(pfreq, 100)\n",
    "plt.title(\"Patient Mutation Frequencies in first cluster\")\n",
    "plt.show()\n",
    "\n",
    "gfreq = [len(c0geneToCases[g]) for g in c0geneToCases]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(gfreq, 100)\n",
    "plt.title(\"Gene Mutation Frequencies in first cluster\")\n",
    "plt.show()\n",
    "\n",
    "print \"Top gene frequencies are \", sorted(gfreq, reverse=True)[0:10]\n",
    "\n",
    "gfreq = [len(c1geneToCases[g]) for g in c1geneToCases]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(gfreq, 100)\n",
    "plt.title(\"Gene Mutation Frequencies in second cluster\")\n",
    "plt.show()\n",
    "\n",
    "print \"number of patients in first cluster \", len(c1patients)\n",
    "\n",
    "print \"Top gene frequencies are \", sorted(gfreq, reverse=True)[0:10]\n",
    "# let's limit to the genes with at least \n",
    "\n",
    "test_minFreq = 10\n",
    "test_genes = [c for c in c0genes if len(c0geneToCases[c]) >= test_minFreq]\n",
    "print \"numbr of genes used is \", len(test_genes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of pairs to test:  94366\n",
      "Number of pairs is  94366\n",
      "number of pairs is  94366\n",
      "Getting cooccurrence across the whole distribution\n",
      "Original pairs  94366 . Min cohort pairs:  94366\n",
      "Writing to file...\n"
     ]
    }
   ],
   "source": [
    "import mutex_triangles as met\n",
    "import chisquared as chi\n",
    "import bingenesbypairs as bgbp\n",
    "compute_mutex = True\n",
    "\n",
    "cpairfile = '/Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-sega-jl-cpairs-min_cohort.txt'\n",
    "test_minFreq = 10\n",
    "\n",
    "\n",
    "genepairs = bgbp.getgenepairs(c0geneToCases, test_genes, test_minFreq=test_minFreq, closer_than_distance=closer_than_distance)\n",
    "print \"Number of pairs is \", len(genepairs)\n",
    "\n",
    "cpairsdict, cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, genepairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(cpairsdict)\n",
    "print \"Getting cooccurrence across the whole distribution\"\n",
    "cpairsdict = chi.add_BinomP_all_pairs(cpairsdict, geneToCases, patientToGenes)\n",
    "\n",
    "cpairsdict = chi.add_BinomP_min_cohort_all_pairs(cpairsdict, geneToCases, patientToGenes, cohort_dict, cohort_dict[0])\n",
    "\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "\n",
    "met.writeanydict(cpairsdict, cpairfile, fieldnames=fieldnames)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We were able to find the genes that cooccur most significantly at a threshold of 14.\n",
    "However they seem to be false positives. We'll need to filter by COSMIC genes, possibly.\n",
    "\n",
    "\n",
    "Gene0\tGene1\tMutationFrequency0\tMutationFrequency1\tBinomProbability\t7SizeMin\t7FreqsMin\t7OverlapMin\t7MinCBinomProb\n",
    "SNORD22loss\tRNU4ATAC3P_MIR548ASloss\t364\t363\t3.34E-70\t95\t[14, 14]\t14\t2.03E-08\n",
    "GPC6_RNA5SP35_DCT_TGDS_GPR180_RNA5SP36_RN7SL585P_LINC00391_SOX21_LINC00557_ABCC4_RNY3P8_RNY4P27_CLDN10_DZIP1_DNAJC3_UGGT2loss\tSNORD22loss\t370\t364\t3.82E-69\t95\t[14, 14]\t14\t2.03E-08\n",
    "SNORD22loss\tGPC5loss\t364\t369\t6.69E-69\t95\t[14, 15]\t14\t4.74E-08\n",
    "HS6ST3_RN7SL164Ploss\tSNORD22loss\t358\t364\t6.86E-68\t95\t[14, 14]\t14\t2.03E-08\n",
    "FAM230Closs\tPEX26_TUBA8_USP18_GGTLC3_PI4KAP1_RN7SKP131_RIMBP3_GGT3P_DGCR6_PRODH_DGCR5_DGCR2_DGCR14_TSSK2_GSC2_SLC25A1_CLTCL1_HIRA_C22orf39_RN7SL168P_MRPL40_UFD1L_CDC45_CLDN5_GP1BB_TBX1_GNB1L_C22orf29_TXNRD2_COMT_MIR4761_ARVCF_TANGO2_MIR185_DGCR8_MIR3618_MIR1306_TRMT2A_RANBP1_ZDHHC8_RTN4R_MIR1286_DGCR6L_FAM230A_USP41_ZNF74_SCARF2_KLHL22_RN7SL812P_MED15_SMPD4P1_POM121L4P_TMEM191A_PI4KA_SERPIND1_SNAP29_CRKL_AIFM3_LZTR1_THAP7_TUBA3FP_P2RX6_SLC7A4_MIR649_P2RX6P_BCRP2_POM121L7_FAM230B_GGT2_RIMBP3B_RN7SKP63_HIC2_TMEM191C_PI4KAP2_RN7SKP221_RIMBP3C_UBE2L3_YDJC_CCDC116_SDF2L1_PPIL2_MIR301B_MIR130B_YPEL1_RN7SL280Ploss\t410\t413\t5.27E-67\t95\t[31, 31]\t31\t6.73E-09\n",
    "SNORD22loss\tLINC00359_RN7SKP7_OXGR1_MBNL2_RNA5SP37_RAP2Aloss\t364\t353\t1.08E-66\t95\t[14, 14]\t14\t2.03E-08\n",
    "SNORD22loss\tLINC00433_LINC00560_LINC00440_LINC00353_LINC00559_RNA5SP34_MIR622_LINC00410_LINC00379_MIR17HGloss\t364\t373\t4.67E-66\t95\t[14, 14]\t14\t2.03E-08\n",
    "SNORD22loss\tIPO5loss\t364\t350\t1.95E-65\t95\t[14, 14]\t14\t2.03E-08\n",
    "SNORD22loss\tFARP1_RNF113B_RN7SKP8_MIR3170_STK24_RN7SL60P_SLC15A1_DOCK9_UBAC2_RN7SKP9_GPR18_GPR183_MIR623_LINC00449_TM9SF2_CLYBL_MIR4306_ZIC5_ZIC2_LINC00554_PCCA_GGACT_TMTC4_LINC00411_NALCN_ITGBL1loss\t364\t362\t6.11E-65\t95\t[14, 15]\t14\t4.74E-08\n",
    "MAPK1_RNA5SP493loss\tFAM230Closs\t405\t410\t3.20E-64\t95\t[30, 31]\t30\t1.30E-08\n",
    "SPRY2loss\tSNORD22loss\t383\t364\t2.64E-57\t95\t[15, 14]\t13\t3.44E-07\n",
    "\n",
    "SNORD22loss position: 11\t62854161\t62854285\n",
    "RNU4ATAC3P_MIR548ASloss position: 13\t92059844\t92059969, 13\t92490163\t92490220\n",
    "GPC6_RNA5SP35_DCT_TGDS_GPR180_RNA5SP36_RN7SL585P_LINC00391_SOX21_LINC00557_ABCC4_RNY3P8_RNY4P27_CLDN10_DZIP1_DNAJC3_UGGT2loss\t position:\n",
    "13\t93226842\t94407401\n",
    "IPO5 loss position: 13\t97953658\t98024297\n",
    "FAM230Closs position: 13\t18195297\t18232024\n",
    "PEX26loss_RN7S2Lloss position: 22\t18077920\t18131138, 22\t21722895\t21723182\n",
    "MAPK1loss: 22\t21754500\t21867680\n",
    "SPRY2loss\t13\t80335976\t80340951\n",
    "\n",
    "Conclusion:\n",
    "## SNORD22 gets legitimate co-occurrence! But with many in same chromosome range: 13, 920000-980000\n",
    "## Explanation: the 14 that cooccur happen to be very highly mutated patients\n",
    "## Solution. Partition over CNA space first, but test on SEG genes\n",
    "\n",
    "\n",
    "## Mutual Exclusivity: nothing seems mutex\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Next plan: use the CNA clusters, but test the segment genes\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ8AAAEnCAYAAAB2e06MAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJHwSiUQiUXf+D2EeSUDK7+FoAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1a35d0090>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0, 1, 2, 3, 4, 5, 6, 7, 8] [0, 1, 2, 3, 4, 5, 6, 7, 8]\n"
     ]
    }
   ],
   "source": [
    "# when we're ready, let's look at the cooccurring pairs here. EDIT PARTITION FILE\n",
    "\n",
    "import mutex as mex\n",
    "import csv\n",
    "\n",
    "mutationmatrix = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-seg-jl.m2'\n",
    "patientFile = '/Users/jlu96/maf/new/BRCA_wustl/shared_patients.plst'\n",
    "geneFile = None\n",
    "minFreq = 0\n",
    "COSMICFile = '/Users/jlu96/conte/jlu/Analyses/CancerGeneAnalysis/COSMIC/COSMICGenes_OnlyLoss.txt'\n",
    "partition_file = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.ppf9'\n",
    "\n",
    "numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "\n",
    "cohort_dict, clusterToProp, min_cohort = load_patient_cohorts(partition_file, patientToGenes)\n",
    "\n",
    "plot_pmm_clusters(patientToGenes, cohort_dict, len(cohort_dict), 100, 'Partitioned BRCA patients in Segments')\n",
    "print clusterToProp.keys(), cohort_dict.keys()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number genes is  488\n",
      "number patients is  80\n",
      "Mean is  243.5\n",
      "Number of genes in cluster 0:  488\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEKCAYAAADkYmWmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2P8HpQw6MPpeGeIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1b11766d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEKCAYAAAAYd05sAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMyuUA97MrFAOeDOzQv1/jp2PnrPaTTQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x15c77f150>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top gene frequencies are  [33, 33, 32, 32, 32, 32, 32, 31, 30, 29]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEKCAYAAADzQPVvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1493f3750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of patients in first cluster  120\n",
      "Top gene frequencies are  [96, 96, 96, 96, 96, 96, 95, 95, 95, 94]\n",
      "numbr of genes used is  38\n"
     ]
    }
   ],
   "source": [
    "c0patients = cohort_dict[0]\n",
    "c1patients = cohort_dict[1]\n",
    "\n",
    "c0genes, c0geneToCases, c0patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c0patients)\n",
    "c1genes, c1geneToCases, c1patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c1patients)\n",
    "\n",
    "print \"number genes is \", len(c0genes)\n",
    "print \"number patients is \", len(c0patients)\n",
    "print \"Mean is \", clusterToProp[0]['Mean']\n",
    "print \"Number of genes in cluster 0: \", len(c0genes)\n",
    "\n",
    "\n",
    "\n",
    "pfreq = [len(c0patientToGenes[p]) for p in c0patients]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(pfreq, 100)\n",
    "plt.title(\"Patient Mutation Frequencies in first cluster\")\n",
    "plt.show()\n",
    "\n",
    "gfreq = [len(c0geneToCases[g]) for g in c0geneToCases]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(gfreq, 100)\n",
    "plt.title(\"Gene Mutation Frequencies in first cluster\")\n",
    "plt.show()\n",
    "\n",
    "print \"Top gene frequencies are \", sorted(gfreq, reverse=True)[0:10]\n",
    "\n",
    "gfreq = [len(c1geneToCases[g]) for g in c1geneToCases]\n",
    "% matplotlib inline\n",
    "plt.figure()\n",
    "plt.hist(gfreq, 100)\n",
    "plt.title(\"Gene Mutation Frequencies in second cluster\")\n",
    "plt.show()\n",
    "\n",
    "print \"number of patients in first cluster \", len(c1patients)\n",
    "\n",
    "print \"Top gene frequencies are \", sorted(gfreq, reverse=True)[0:10]\n",
    "# let's limit to the genes with at least \n",
    "\n",
    "test_minFreq = 5\n",
    "test_genes = [c for c in c0genes if len(c0geneToCases[c]) >= test_minFreq]\n",
    "print \"numbr of genes used is \", len(test_genes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['CNEP1R1_HEATR3_RNY4P3_PAPD5_ADCY7_BRD7_NKD1_SNX20_NOD2_CYLD_RNA5SP426loss', 'CHD9_RNA5SP427_RBL2_AKTIPloss', 'GPT2_DNAJA2loss', 'ANKRD26P1loss', 'HNRNPA1P48_RN7SKP142loss', 'GNAO1_MIR3935_AMFR_NUDT21_OGFOD1_BBS2_MT4_MT3_MT2A_MT1L_MT1E_MT1M_MT1JP_MT1A_MT1DP_MT1B_MT1F_MT1G_MT1H_MT1X_NUP93_SLC12A3_HERPUD1_CETP_NLRC5_CPNE2_FAM192A_RSPRY1_ARL2BP_PLLP_CCL22_CX3CL1_CCL17_CIAPIN1_COQ9_POLR2C_DOK4_CCDC102Aloss', 'RN7SL134P_MIR3182_HSBP1_MLYCD_OSGIN1_NECAB2_SLC38A8_RNA5SP432_MBTPS1_HSDL1_DNAAF1_TAF1C_ADAD2_KCNG4_RNA5SP433_WFDC1_ATP2C2_TLDC1_COTL1_KLHL36_USP10_CRISPLD2_ZDHHC7_KIAA0513_FAM92B_LINC00311_MIR5093loss', 'GSE1_RN7SL381P_GINS2_C16orf74_EMC8_COX4I1_IRF8_LINC00917_FENDRR_FOXF1_MTHFSD_FOXC2_FOXL1loss', 'ABCC11_LONP2_SIAH1_N4BP1loss', 'ABCC12loss']\n"
     ]
    }
   ],
   "source": [
    "print test_genes[0:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of pairs to test:  2\n",
      "Number of pairs is  2\n",
      "number of pairs is  2\n",
      "Getting cooccurrence across the whole distribution\n",
      "Original pairs  2 . Min cohort pairs:  2\n",
      "Writing to file...\n"
     ]
    }
   ],
   "source": [
    "import mutex_triangles as met\n",
    "import chisquared as chi\n",
    "import bingenesbypairs as bgbp\n",
    "compute_mutex = True\n",
    "\n",
    "closer_than_distance = 50000000\n",
    "cpairfile = '/Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-seg-jl-cpairs-min_cohort_cna_partitions.txt'\n",
    "test_minFreq = 5\n",
    "\n",
    "\n",
    "genepairs = bgbp.getgenepairs(c0geneToCases, test_genes, test_minFreq=test_minFreq, closer_than_distance=closer_than_distance)\n",
    "print \"Number of pairs is \", len(genepairs)\n",
    "\n",
    "cpairsdict, cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, genepairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(cpairsdict)\n",
    "print \"Getting cooccurrence across the whole distribution\"\n",
    "cpairsdict = chi.add_BinomP_all_pairs(cpairsdict, geneToCases, patientToGenes)\n",
    "\n",
    "cpairsdict = chi.add_BinomP_min_cohort_all_pairs(cpairsdict, geneToCases, patientToGenes, cohort_dict, cohort_dict[0])\n",
    "\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "\n",
    "met.writeanydict(cpairsdict, cpairfile, fieldnames=fieldnames)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# All found pairs were within 100 MB of each other\n",
    "120 pairs, 30MB away from each other\n",
    "388 pairs > 10 MB from each other\n",
    "\n",
    "Found pairs:\n",
    "\n",
    "ZNF267_HERC2P4_TP53TG3D_HERC2P5_SLC6A10P_HERC2P8_TP53TG3C_TP53TG3B_TP53TG3_ARHGAP23P1_LINC00273_RNA5SP406_RNA5SP407_RNA5SP408_RNA5SP409_RNA5SP410_RNA5SP413_RNA5SP415_RNA5SP416_RNA5SP417_RNA5SP418_RNA5SP419_RNA5SP420_RNA5SP421_RNA5SP422_RNA5SP423loss\n",
    "16\t31873758\t31917357 to 16\t35755177\t35755285\n",
    "C16orf95_FBXO31_MAP1LC3B_ZCCHC14_JPH3_KLHDC4_SLC7A5_CA5A_BANPloss\n",
    "16\t87083562\t87317420 to 16\t87949244\t88118422\n",
    "\n",
    "No COSMIC genes found\n",
    "P = 0.0007\n",
    "\n",
    "Coocurrs in all of them, p < 0.15 in most"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[frozenset(['ZNF267_HERC2P4_TP53TG3D_HERC2P5_SLC6A10P_HERC2P8_TP53TG3C_TP53TG3B_TP53TG3_ARHGAP23P1_LINC00273_RNA5SP406_RNA5SP407_RNA5SP408_RNA5SP409_RNA5SP410_RNA5SP413_RNA5SP415_RNA5SP416_RNA5SP417_RNA5SP418_RNA5SP419_RNA5SP420_RNA5SP421_RNA5SP422_RNA5SP423loss', 'C16orf95_FBXO31_MAP1LC3B_ZCCHC14_JPH3_KLHDC4_SLC7A5_CA5A_BANPloss']), frozenset(['ZNF469_ZFPM1_MIR5189_ZC3H18_IL17C_CYBA_MVD_SNAI3_RNF166_CTU2_PIEZO1_MIR4722_CDT1_APRT_GALNS_TRAPPC2L_PABPN1L_CBFA2T3_ACSF3_LINC00304_CDH15_SLC22A31_ZNF778_ANKRD11_SPG7_RPL13_SNORD68_CPNE7_DPEP1_CHMP1A_SPATA33_CDK10_SPATA2L_VPS9D1_ZNF276_FANCA_SPIRE2_TCF25_MC1R_TUBB3_DEF8_CENPBD1_AFG3L1P_DBNDD1_GAS8_URAHP_PRDM7_TUBB8P7_FAM157Closs', 'ZNF267_HERC2P4_TP53TG3D_HERC2P5_SLC6A10P_HERC2P8_TP53TG3C_TP53TG3B_TP53TG3_ARHGAP23P1_LINC00273_RNA5SP406_RNA5SP407_RNA5SP408_RNA5SP409_RNA5SP410_RNA5SP413_RNA5SP415_RNA5SP416_RNA5SP417_RNA5SP418_RNA5SP419_RNA5SP420_RNA5SP421_RNA5SP422_RNA5SP423loss'])]\n",
      "16\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEKCAYAAADkYmWmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl8VOXd9/HPL2GJLCYCFgRCplLX1oLeyI0i3qmKVUCl\nm4CIBLF9HosKIm6gEhBEW6tQrbY3IihFsXZBpKjVyvhopU1pBa0WLEjYwpaEsElYkuv545yEScgy\nmSTMHPJ9v155Zc4517nO72y/OXOdM3OZcw4REQmmpHgHICIisVMSFxEJMCVxEZEAUxIXEQkwJXER\nkQBTEhcRCTAlcTmGmXUzs71mZvGOpT7MrNTMTo/Tss8ys5VmtsfMbotHDI0hSMeGmS01sxHxjqOx\nBSKJm9klZvahmRWZWYGZfWBmveIdV12Y2Twze7iWMqVmtt3MkiPGNTezHWZWGuVyQn49Ue9bM8s1\ns8vKhp1zG51zbZ2+RFAf9wB/ds6d7Jx7Ot7BNJREPTbMLNvM5keOc84NcM7Nr26eOtSdZWbv17ee\nxpLwSdzMTgaWALOAU4AuwBTgYDzjakSFwNURw1f74+p60tTlSsnVsXyTYmbNYpgtA/isoWOJpxi3\ngzQ251xC/wG9gF21lLkZ74QpBN4EukVMuxJYAxQBvwDeA0b707KAvwBPALuAtcDFwChgI7AduCmi\nrpbA48AGYBvwLJDiT8sENgPj/fnygCx/2o+AQ3hvPHuB16pZj1JgIvCbiHG/9ceVRozLBS6PGM4G\n5vuvN/r17PX//hvoDrwL5AM7gV8DqX75+UAJ8KVffgIQ8utI8st0BhYDBcB/gFsqLfs3wAvAHuBf\nwH/VsK9Kgf8DfO5v86erWg9/uHIcYeBhf5/t9WPqACwAdgM5QEalZd0OrPPX+yeARXnclAI/9td3\nXTXrci3wqb8ey4Cz/fHvAkeAA/42+VoV87YD5gJb/OX/IWLaD/3lFgCvAadFTLsY+Dve8ZwDXBRl\nndcBK/3ttBa4MopjqWz734x3zIfx3pwq75OpwAf+ur4FtI+o7yZ/3nzggcrLq7RN5gG/BP7k1xWu\ntE9m4R3fu4EVwCX++Kvwzq1D/nHxUURso+uwv485LoFz/P14xK+70B8/wN/3e/DO+7viliPjteCo\nA4S2/gEwz99Zp1Safp1/wJ+F98liEvAXf1oHf4cP9qfd4e/om/3pWcBhYCTelejD/g55CmgO9Pd3\nUiu//JPAIiANaIOXRB7xp2X6dWUDyXhX0Ps5miznAlNrWddS4Ot4bxAn433y2OaPi0zi64HLIoYn\nc/TEq3CS+eO6A5f769QB743syRrqC1HxRP1/wNNAC6AHsAP4VsRJf8DfNwY8AiyvZR0X++uX7tf1\n7crrUU0cYbyT7Kv+/J/6+/4yf5u/ADxfaVl/9vdXOt6bedkbeLXHTcS8b/nztqxiPc4E9vnbNRm4\n26+vmT99Gf5xVs12+CPwMpAKNAP6+eMvw3vD6elv758D7/nT2uElmOF+zEPxEtIptdTZGy/pX+4P\ndwbOiuJYKtv+84CT8C5iqton/wG+BqT46z3Dn3YuXuK7GO/Y+yne+XdZNdtkHt75dom/7jOB9yOm\nD8c7J5LwLpa2Ai0i4n6xUn3l+yDK/V3dcTkyMg5/3Fagr/86FTg/bjkyXguuU5BwNl4S3ISXKF8D\nvuJPeyPyZPF30H6gG95VwF8q1bWRikn884hp5/k789SIcfnAN/ES1D7g9IhpFwFf+K8z8a5mI5Pn\ndqC3/3ou8HAt61mKl3Bn4129/1/gV/64mpJ4NseeeEk1LGcw8M8a6iuvwz+gjwCtI6Y/AsyNWPaf\nIqadC3xZyzpeHDH8CnBv5fWoal3wTsr7I6Y/DvwxYngQ/lVYxLKujBi+FXinluMmPWLezBrW40Fg\nYcSw4V0AXBoR6+hq5j0N79NPahXT5gCPRgy3xkt8GcAI4K+Vyn+Il2RqqvNXwM+qiSWaYylUyz6Z\nWGkbv+G/fghYEDHtJLwr5pqS+EuV1v0I0KWa8oXAeVUdOxGxlZ3r0ezv6o7LLI5N4hvwztGTazqn\nj8dfwreJAzjnVjvnRjnn0oFv4F1JzPQnZwCzzGyXme3C+wgKXtv5aXgnVqTKw9sjXh/wl7ez0rg2\nwKlAK+AfEct6A+/KtkyBcy7yBuSX/rx14YAX8U7MEf7rerVXm1lHM1toZpvNbDdeE0r7KGfvjPcR\ncn/EuI1427dM5Db8Ekip5cbqtkrlW0cZS+VlFeNdMUUOV97emyJeb8RbH6j5uKlq3spO8+sDwHln\n9qZK87tq5k3H26a7q6l3Q0S9+/3Yyo7njZXKb/DXqWsNdXbFa1KKVU3bASruz7LzBT+u8vPNOXeA\no9u5Kq5S+f14ibozgJlNMLPP/AccduFdAXeosqZjRbO/63Jcfg+vSSXXzMJm1ifKOBpcIJJ4JOfc\nGryPzd/wR20EfuScOyXir7VzbjneR56uZfP6j0V1PabS6OTjHaDnRiwnzTl3crShR7sg59z7QCe8\nTxt/qaLIfioeYJ1qWc4jeFdp33DOpeK9OSTVMk+ZPKCdmUUmx24c+2bYEPbjvVGW6VRdQV8027Rb\npddb/NfVHTd/jbL+PLzEAJQfW+kR9ddkE942Ta2m3lBEva3x3nA3V16mL8NfZk11bsJr7qhKTcdS\nmaiP3UryqHj+nUTNFw9l27CsfBu8JqQ8M+uH12T1A/+8OwWvqbTsAqe2GKPZ39U5pm7n3Arn3GC8\ni7tFePeF4iLhk7j/vO14M+viD6cDw4DlfpFfAhPN7Fx/eqqZ/cCfthQ4z8yu8++sj6H2xFAl/wp7\nNjDTzE71l9XFzK6MsortQF2eWb4G78ZZVVYCQ82smf+o5fc4eqDt5GizTJk2eCfrHn873l1FbN2p\ngnNuE95H9hlm1tLMvol3g+jXdViXmhhHT8SVwKVmlu4no/urKV/V6+pMMLM0/7i5A+9jMtR83ETj\nN8BAM7vMzJoDd+F9Eviwtvicc1vxPsU948fW3Mwu9Se/DIwysx5m1hLvDfivzrmN/jxnmtkwf98P\nwWtqXOKc21ZDnXP8Oi8zsyT/uD3Ln1bTsRSt6vbD74BrzOwiM2uB1+RR2z4bYGZ9/fIP491f2YJ3\nb+wIkG9mLczsIbz26zLbgFANz6/XdX9HHpfbga7+fi577He4maU650rw2v1LalmvRpPwSZyjT1j8\nzcz24SXvj/FOGpxzi4DHgIV+U8EnwLf9afnAD/CeSsjHu9O8gqOPJzqOPWBrOoDvxbuz/1d/WW/j\n3eCKZt45wLn+x7nfV1OmfH7n3GfOuX9XU/eDeEl3F96JsSBivi+B6cBfzKzQzHrjPZJ5Ad6Vy+t4\nJ1dkfTOAB/zYxlexvGF4V4d5wO+Bh5xz70aUq8s2rKqs82N/Gy/Jfoz3BMbrtdQdzbJfA/4BfIT3\nqOrz/rKqPW6iWAecc58DN+LdBN8JDASucc4dibKOEXj3d1bjJYk7/Hr/jLd/f4e3vb+KdwMT51wB\nXrv/XXjH8wRgkHOusJY6/473xNWTeDc4wxz9hFLtsVTDOkS1T5xzn+I9HbTQX5e9eM1f1T0e7ICX\n8G5SFgDn421j8J4meRPvxnYu3qfiyKalV/3/BWa24piK676/I4+tP+PdRN9mZmXNdzcC6/26foR3\n0zUuzG+kr3qi2fN4B+cO59x5/rh2eCdaBt7GvN45V9T4odaf3067CbjBOfdevOMRaUr85pFdeI9c\nbqhi+lxgs3PuweMeXIDVdiU+F+/RsUj3AW87587Ee4e6rzECayhmdqX/8bIl3vPWANG0g4lIPZnZ\nNWbWym/bfxz4uKoEXlb8OIZ2wqgxifs32HZVGn0t3o1F/P+DGyGuhnQRXhNI2Ufewc65E/XbniKJ\n5lq8G69b8JpthtZQtqrmMalFjc0p4P0WB/B6RHPKLv/OcNkd+cKyYREROb7qdWPTfzZW75wiInES\nyw/abDezTs65bWZ2GhW/bFHOzJTcRURi4JyL+v5ALFfii/G+TYj/f1F1BScvm8wDf3qAH074YVy/\nllrT3+TJk+Meg+JUjIpTcZb91VWNSdzMXsb78sJZZrbJzEYBjwL9zexzvB/rebTOSxURkQZRY3OK\nc25YNZOuaIRYRESkjoLwjc1GlZmZGe8QoqI4G04QYgTF2dCCEmdd1fqIYcwVm7nJyyZTcriE7X/a\nzv/+9H8bZTkiIicSM8PV4camulsSaQTV/w6TyFENcRGtJC7SSBrrU66cGBrqjb7Jt4mLiASZkriI\nSIApiYuIBJiSuIgE2q233sq0adPiHUbc6MamyHEyblw2RY3YfUpaGsycmR1V2ZdeeoknnniCNWvW\n0LZtW3r27MmkSZPo27dv4wXYSJ599tl4hxBXSuIix0lREYRC2Y1Wf25udHU/8cQTPPbYY/zqV7/i\n29/+Ni1atODNN99k8eLFCZ3ES0pKSE5OrjCutLSUpKSm3aDQtNdepInZvXs3kydP5plnnmHw4MGc\ndNJJJCcnM3DgQB577DEOHjzIuHHj6NKlC126dOHOO+/k0KFD5fPPnj2bM844g/bt23PdddexdevW\n8mmffvop/fv3p3379nTq1IkZM2YAXvJ95JFH+NrXvsbJJ59Mr1692LJlC7m5uSQlJVFaWlpeR2Zm\nJnPmzAFg3rx59O3bl/Hjx9OhQweys7MZNWoUt956KwMGDKBNmzYsW7aMrKwsHnzQ69EtHA7TtWtX\nnnjiCTp27Ejnzp2ZN29eef0FBQVcc801pKam0rt3bx544AH69evXmJu80SmJizQhy5cvp7i4mO98\n5ztVTp8+fTo5OTmsWrWKVatWkZOTU97e/O677zJx4kReffVVtm7dSkZGBkOHeh317N27lyuuuIIB\nAwawdetW1q5dy+WXXw54V/4LFy7kjTfeYM+ePcydO5eTTjqpyuWbWYXnp3NycujevTs7duxg0qRJ\nOOd4+eWXefDBB9m3bx+XXHLJMfNs376dPXv2kJeXx5w5cxgzZgy7d+8GYMyYMbRt25bt27fzwgsv\n8OKLLwb+i1lK4iJNSEFBAR06dKi2CeKll17ioYceokOHDnTo0IHJkyczf/58ABYsWMDo0aPp2bMn\nLVq0YMaMGSxfvpwNGzawZMkSOnfuzJ133kmLFi1o06YNvXv3BmDOnDlMnz6dM844A4DzzjuPdu3a\nRRVv586dGTNmDElJSaSkpGBmDB48mIsuugiAli1bAhW/WNW8eXMeeughkpOTufrqq2nTpg1r1qyh\npKSE3//+90yZMoWUlBTOOeccRo4cGfgvZSmJizQh7du3Jz8/v0ITRqS8vDwyMjLKh7t160ZeXh5A\n+dV3mdatW9O+fXu2bNnC5s2bOf3006usc9OmTXTv3j2meNPT06MaF6l9+/YV3qRatWrFvn372Llz\nJ0eOHKkwf9euXWOKK5EoiYs0IRdddBEtW7bkD3/4Q5XTO3fuTG5ubvnwxo0b6dKlS5XT9u/fT0FB\nAV27diU9PZ0vvviiyjrT09NZu3btMeNbt24NwJdfflk+btu2bRXKRNvUEU25U089lWbNmrFp06by\ncZGvg0pJXKQJSU1NZerUqYwZM4bXXnuNL7/8ksOHD/PGG29w7733MmzYMKZNm0Z+fj75+flMnTqV\nG2+8EYBhw4Yxd+5cVq1axcGDB5k4cSJ9+vShW7duDBw4kK1btzJr1iwOHjzI3r17ycnJAeCWW27h\nwQcfZO3atTjn+PjjjyksLOTUU0+lS5cuzJ8/n5KSEp5//nnWrVtXY/xVNX1E2yNOcnIy3/3ud8nO\nzubAgQOsXr2a+fPnB75NXI8YihwnaWnRPwYYa/3RGD9+PJ06dWLatGkMHz6ctm3b0qtXLyZNmsT5\n55/Pnj17+OY3vwnA9ddfzwMPPADA5ZdfzsMPP8z3vvc9du3aRd++fVm4cCEAbdu25e2332bs2LFM\nmTKFli1bcuedd9K7d2/Gjx/PwYMHufLKK8nPz+ecc84p/yQwe/ZsfvzjHzNx4kRGjx5d4RHHyjcs\nox1XU1J++umnycrKolOnTpx99tkMGzaMFStWRLfhEpR+T1ykEfi/CR3vMKQW9957Lzt27GDu3LnH\nfdnVHSN1/T1xNaeISJOxZs0aPv74Y5xz5OTk8Pzzz1f7uGVQqDlFRJqMvXv3MmzYMPLy8ujYsSMT\nJkzg2muvjXdY9aIkLiJNRq9evfjPf/4T7zAalJpTREQCTElcRCTAlMRFRAJMSVxEJMCUxEVEAkxJ\nXESkBgMGDCj/JcdEpEcMRY6TcfeNo6i48fpnS0tJY+ajM2stFwqFOHDgAOvXr6dVq1YAPPfccyxY\nsIBly5bVOn9mZiYjRoxg9OjR9Y45CJYuXRrvEGqkJC5ynBQVFxEaHGq0+nMX5UZdtrS0lFmzZnH/\n/ffXeTlB/8Go6lTu/q3sK/GJvr5qThFpYsyMCRMm8Pjjj5f3eFPZhx9+yIUXXkhaWhq9e/dm+fLl\nAEyaNIn333+f2267jbZt23LHHXdUOf8HH3zAxRdfzCmnnEK3bt144YUXAK97uJtuuomvfOUrhEIh\npk+fXp4snXNMmzaNUChEx44dGTlyJHv27Km1zgMHDnDXXXcRCoVIS0ujX79+FBcXEw6Hj/nt8VAo\nxLvvvgtAdnY23//+9xkxYgSpqanMmzePzMzM8g6j27RpwxdffHFMl3GXXHIJd999N+3ateP000/n\nzTffLK9//fr1XHrppZx88sn079+fMWPGMGLEiDrvo7pQEhdpgnr16kVmZiaPP/74MdMKCwsZOHAg\n48aNo7CwkPHjxzNw4EB27drF9OnT6devH7/4xS/Yu3cvP//5z4+Zf8OGDQwYMICxY8eSn5/PypUr\n6dmzJwC33347e/fuZf369bz33nu8+OKL5T8+NXfuXF544QXC4TBffPEF+/bt47bbbqu1zgkTJvDR\nRx+xfPlyCgsL+elPf1ptz0WVr6oXL17MD37wA3bv3s3w4cMBrwej5557jr1795KRkVFll3Fnn302\nBQUF3HPPPRWalW644Qb69OlDYWEh2dnZ/PrXv270K3klcZEmyMyYOnUqTz31FPn5+RWm/fGPf+Ss\ns85i+PDhJCUlMXToUM4++2wWL15cXqamX2h86aWX6N+/P0OGDCE5OZl27drRo0cPSkpKeOWVV5gx\nYwatW7cmIyODu+66q0L3b2VX1K1bt2bGjBksXLiQkpKSaussLS1l7ty5zJo1i9NOO42kpCT69OlD\nixYtotoOF198cflvp5R1/5aVlcU555xDUlISzZod2+KckZHB6NGjMTNuuukmtm7dyo4dO9i4cSMr\nVqxg6tSpNGvWjL59+3Lttdc2+q9ZKomLNFFf//rXGTRoEI8++miFq8W8vDy6detWoWxGRkZ5N21Q\ncztxdV215efnc/jw4WO6f9uyZQtwbPdv3bp148iRI2zfvr3GOouLi2Pu/q2q7tlq6/6tU6dO5a/L\nbgzv27ePvLw82rVrR0pKStR1NQQlcZEmbMqUKcyePbs8kQJ06dKFDRs2VCi3YcOG8m7aamseSE9P\nr7KHng4dOtC8efNjun8rS6RVdQ3XrFkzOnXqVGOdKSkp1Xb/Ftn1W0lJCTt37qxQpqp1ibX547TT\nTqOwsJADBw5UWIfGpiQu0oR1796dIUOGMGvWrPJxV199NZ9//jkvv/wyR44c4ZVXXmH16tUMGjQI\ngI4dO9bYjdrw4cN55513ePXVVzly5AgFBQWsWrWK5ORkrr/+eiZNmsS+ffvYsGEDTz75ZIXu3558\n8klyc3PZt28fEydOZOjQoSQlJXHDDTdUWWdSUhI333wz48ePZ+vWrZSUlLB8+XIOHTrEmWeeSXFx\nMUuXLuXw4cNMmzaNgwcP1rpNYm3+yMjIoFevXmRnZ3P48GGWL1/OkiVLGr1NXI8YihwnaSlpdXoM\nMJb6Y/HQQw9V6Guyffv2LFmyhLFjx3LrrbdyxhlnsGTJEtq1awfA2LFjGTlyJM8++yw33XQTM2dW\nfDY9PT2dpUuXMmHCBG655RZSU1OZPn06PXr04KmnnuL222/n9NNPJyUlhR/96EeMGjUKgJtvvpm8\nvDwuvfRSiouLueqqq3jqqacAr2mlujoff/xx7r//fi688EL27dtHz549eeutt0hNTeWZZ57hlltu\noaSkhHvuuadC80ZVXb2Vja9Odd3DlVmwYAFZWVm0b9+e3r17M2TIEEpKSuqyO+os5u7ZzOx+4Eag\nFPgEGOWcOxgxXd2zSZOl7tkEYMiQIZx77rlMnjz5mGlx7Z7NzELAD4ELnHPnAcnA0FjqEhE5UaxY\nsYJ169ZRWlrKG2+8weLFixk8eHCjLjPW5pQ9wGGglZmVAK2ALTXPIiJyYtu2bRvf/e53KSgoID09\nnV/+8pf06NGjUZcZUxJ3zhWa2c+AjcAB4C3n3DsNGpmISMAMGjSo/Abw8RJTEjez7sA4IATsBl41\ns+HOuQWR5cLzwpSWlLJ/3X7C4TCZmZn1jVdE5IQSDocJh8Mxzx/TjU0zGwL0d87d4g+PAPo458ZE\nlNGNTWmydGNTahPXG5vAaqCPmZ1k3vM1VwCfxViXiIjEKNY28VVm9iKwAu8Rw38CutQWiZDoP2Eq\nJ4aYv+zjnPsJ8JMGjEXkhKGmFDle9LV7EZEAUxIXEQkwJXERkQBTEhcRCTAlcRGRAFMSFxEJMCVx\nEZEAUxIXEQkwJXERkQBTEhcRCTAlcRGRAFMSFxEJMCVxEZEAUxIXEQkwJXERkQBTEhcRCTAlcRGR\nAFMSFxEJsJi7Z6uLDz/8kKxxWeXDaSlpzHx05vFYtIjICe24JPFDdojQ4FD5cO6i3OOxWBGRE56a\nU0REAkxJXEQkwJTERUQCTElcRCTAlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRF\nRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQCLOYkbmZpZvZbM/u3mX1mZn0aMjAREaldfTqFmAUsdc59\n38yaAa0bKCYREYlSTEnczFKBfs65kQDOuSPA7oYMTEREahdrc8pXgZ1mNtfM/mlms82sVUMGJiIi\ntYs1iTcDLgCecc5dAOwH7muwqEREJCqxtolvBjY75/7uD/+WKpJ4eF6Y0pJSCjYWkLsyl1DPUIyL\nE2kY48ZlU1TkvU5Lg5kzs+Maj0g4HCYcDsc8f0xJ3Dm3zcw2mdmZzrnPgSuATyuXy8zKpORwCdvW\nb1MCl4RQVAShUDYAubnZcY1FBCAzM5PMzMzy4SlTptRp/vo8nXI7sMDMWgDrgFH1qEtERGIQcxJ3\nzq0CLmzAWEREpI70jU0RkQBTEhcRCTAlcRGRAFMSFxEJMCVxEZEAUxIXEQkwJXERkQBTEhcRCTAl\ncRGRAFMSFxEJMCVxEZEAUxIXEQkwJXERkQBTEhcRCTAlcRGRAFMSFxEJMCVxEZEAUxIXEQmw+vSx\nGbOcv+WQNS4LgLSUNGY+OjMeYUjAjLtvHEXFXlf1Om5EPHFJ4ofsEKHBIQByF+XGIwQJoKLiIh03\nIpWoOUVEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQC\nTElcRCTAlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQCrF5J3MySzewjM3u9oQISEZHo1fdKfCzw\nGeAaIBYREamjmJO4mXUFBgDPAdZgEYmISNTqcyX+JHA3UNpAsYiISB3FlMTNbBCwwzn3EboKFxGJ\nm1h7u78YuNbMBgApwMlm9qJz7qbIQuF5YUpLSinYWEDuylxCPUP1DFekduPGZVNU5L1OS4OZM7Pj\nGo9ITcLhMOFwOOb5Y0rizrmJwEQAM/sfYELlBA6QmZVJyeEStq3fpgQux01REYRC2QDk5mbHNRaR\n2mRmZpKZmVk+PGXKlDrN31DPievpFBGROIi1OaWcc+494L0GiEVEROpI39gUEQkwJXERkQBTEhcR\nCTAlcRGRAFMSFxEJMCVxEZEAUxIXEQkwJXERkQBTEhcRCTAlcRGRAFMSFxEJMCVxEZEAUxIXEQkw\nJXERkQBTEhcRCTAlcRGRAFMSFxEJMCVxEZEAq3f3bMfLuPvGUVRcVD6clpLGzEdn1lq2pnJyYsr5\n+E1W5uYCsPHzZWRlRUzLWUkoFI+oRBpHYJJ4UXERocGh8uHcRblRla2pnJyYDiUV85XMEADr1jnK\ner4H+OCDwfEJSqSRqDlFRCTAlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRFRAJM\nSVxEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRFRAJMSVxEJMCUxEVEAiymJG5m6Wa2zMw+NbN/\nmdkdDR2YiIjULtaefQ4DdzrnVppZG+AfZva2c+7fDRibiIjUIqYrcefcNufcSv/1PuDfQOeGDExE\nRGpX7zZxMwsB5wN/q29dIiJSN/XqKNlvSvktMNa/Iq8gPC9MaUkpBRsLyF2ZS6hnqD6LqyDnbzlk\njcsC1KP9uPvGUVRcVD7cFLZH5P6H2NY5J+evZGVlH60jDWbOzD6mXENs36a4jyQ64XCYcDgc8/wx\nJ3Ezaw78Dvi1c25RVWUyszIpOVzCtvXbGjSBAxyyQ+rR3ldUXFS+LaBpbI/I/Q+xrfOhQymEQtlH\n68jNrrJcQ2zfpriPJDqZmZlkZmaWD0+ZMqVO88f6dIoBc4DPnHO6nBARiZNY28T7AjcC3zKzj/y/\nqxowLhERiUJMzSnOuQ/QF4VEROJOiVhEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRFRAJMSVxE\nJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTA\n6tXbfUOo3Gv5qhWr6NGrBxB9j+CV68j5R06FTmnLVO5xvLpl1aVn8siykfVFvq6p/sp1xzqtOg3d\nU3td5o91vsa0c+8nLApnlb+ORk7OSlaSWz58KKeo+sKNqCH2pUQnSNs67km8cq/lH+R8UOde7Kuq\noyqVexzSjd9dAAAFk0lEQVSvbll16Zk8smxkfZGva6q/ct2xTosmvrrMV9/l1me+xlTa4jBpmSHv\n9bp3o5rn0CH4Slpm+fDmQ4saIbLaNcS+lOgEaVurOUVEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTA\nlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTERUQCTElcRCTAlMRFRAJMSVxEJMCUxEVEAkxJXEQkwJTE\nRUQCLOYkbmZXmdlqM/uPmd3bkEGJiEh0YkriZpYMPA1cBZwLDDOzcxoysOMld2VuvEOISjgcjncI\nUQlCnEW5ufEOISpB2JagOOMt1ivx3sBa51yuc+4wsBC4ruHCOn6UxBtWEOJUEm9YijO+Yk3iXYBN\nEcOb/XEiInIcxdpRsoumUPjFMK7UceTQkRgXIyIiNTHnosrHFWcy6wNkO+eu8ofvB0qdc49FlKl7\nxSIignPOoi0baxJvBqwBLgfygBxgmHPu33WuTEREYhZTc4pz7oiZ3Qa8BSQDc5TARUSOv5iuxEVE\nJDE0yjc2E/GLQGb2vJltN7NPIsa1M7O3zexzM/uTmaXFM0Y/pnQzW2Zmn5rZv8zsjkSM1cxSzOxv\nZrbSzD4zsxmJGGcZM0s2s4/M7HV/OOHiNLNcM/vYjzMngeNMM7Pfmtm//X3/34kUp5md5W/Dsr/d\nZnZHIsUYEev9/rn+iZm9ZGYt6xpngyfxBP4i0Fy8mCLdB7ztnDsT+LM/HG+HgTudc18H+gBj/O2X\nULE654qBbznnegLfBL5lZpeQYHFGGAt8xtEnqxIxTgdkOufOd8719sclYpyzgKXOuXPw9v1qEihO\n59wafxueD/wX8CXwh0SKEcDMQsAPgQucc+fhNU0Ppa5xOuca9A+4CHgzYvg+4L6GXk6MsYWATyKG\nVwMd/dedgNXxjrGKmBcBVyRyrEAr4O/A1xMxTqAr8A7wLeD1RN33wHqgfaVxCRUnkAp8UcX4hIoz\nIq4rgfcTMUagHd4DIqfg3Z98Hehf1zgbozklSF8E6uic2+6/3g50jGcwlfnv1OcDfyMBYzWzJDNb\n6cezzDn3KQkYJ/AkcDdQGjEuEeN0wDtmtsLMfuiPS7Q4vwrsNLO5ZvZPM5ttZq1JvDjLDAVe9l8n\nVIzOuULgZ8BGvKf8ipxzb1PHOBsjiQfyTqnz3vYSJnYzawP8DhjrnNsbOS1RYnXOlTqvOaUrcKmZ\nfavS9LjHaWaDgB3OuY+AKp+9TYQ4fX2d1wRwNV4zWr/IiQkSZzPgAuAZ59wFwH4qfdxPkDgxsxbA\nNcCrlaclQoxm1h0Yh9dC0BloY2Y3RpaJJs7GSOJbgPSI4XS8q/FEtN3MOgGY2WnAjjjHA4CZNcdL\n4POdc4v80QkZK4BzbjfwR7z2x0SL82LgWjNbj3dFdpmZzSfx4sQ5t9X/vxOvDbc3iRfnZmCzc+7v\n/vBv8ZL6tgSLE7w3w3/42xMSb1v2Aj50zhU4544Av8drjq7TtmyMJL4COMPMQv474RBgcSMspyEs\nBkb6r0fitT/HlZkZMAf4zDk3M2JSQsVqZh3K7pqb2Ul4bXkfkWBxOucmOufSnXNfxfto/a5zbgQJ\nFqeZtTKztv7r1nhtuZ+QYHE657YBm8zsTH/UFcCneO25CROnbxhHm1IgwbYlXtt3HzM7yT/vr8C7\n+V63bdlIDfZX4zXYrwXuj+fNg4iYXsZrdzqE12Y/Cu/GwjvA58CfgLQEiPMSvLbblXhJ8SO8p2oS\nKlbgPOCffpwfA3f74xMqzkox/w+wOBHjxGtrXun//avsvEm0OP2YeuDdyF6Fd/WYmmhxAq2BfKBt\nxLiEitGP6R68N8FPgBeA5nWNU1/2EREJMHXPJiISYEriIiIBpiQuIhJgSuIiIgGmJC4iEmBK4iIi\nAaYkLiISYEriIiIB9v8Bx74wONr31DEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1094454d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEKCAYAAADticXcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x19c48d8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print cpairsdict.keys()\n",
    "\n",
    "cooccurpatients = [patient for patient in c0patients if len(patientToGenes[patient].intersection(set(cpairsdict.keys()[0]))) >= 2]\n",
    "missingpatients = c0patients.difference(cooccurpatients)\n",
    "print len(cooccurpatients)\n",
    "\n",
    "\n",
    "# _, _, _, _, _, old_cna_pTG = mex.load_mutation_data('/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.m2', patientFile, geneFile, minFreq)\n",
    "\n",
    "seg_mutations = [len(patientToGenes[p]) for p in cooccurpatients]\n",
    "seg_mutations_miss = [len(patientToGenes[p]) for p in missingpatients]\n",
    "cna_mutations = [len(old_cna_pTG[p]) for p in cooccurpatients]\n",
    "cna_mutations_miss = [len(old_cna_pTG[p]) for p in missingpatients]\n",
    "\n",
    "% matplotlib inline\n",
    "plt.title(\"Segment Mutation number of cooccuring patients\")\n",
    "bins = np.linspace(0, 80, 100)\n",
    "plt.hist(seg_mutations, alpha=0.5, label='Cooccurring', bins=bins)\n",
    "plt.hist(seg_mutations_miss, alpha=0.5, label=\"Not cooccurring\", bins=bins)\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "bins = np.linspace(0, 600, 100)\n",
    "plt.title(\"CNA Mutation number of cooccuring patients\")\n",
    "plt.hist(cna_mutations, alpha=0.5, label='Cooccurring', bins=bins)\n",
    "plt.hist(cna_mutations_miss, alpha=0.5, label=\"Not cooccurring\", bins=bins)\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ8AAAEnCAYAAAB2e06MAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ns5PmdtcTn3cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1083b60d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "488\n"
     ]
    }
   ],
   "source": [
    "only0 = {0: cohort_dict[0]}\n",
    "plot_pmm_clusters(c0patientToGenes, only0, 1, 20, 'Partitioned BRCA patients in Segments')\n",
    "print len(c0genes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  We're going to take the mutations in the first cluster, and try testing pairs across all the clusters\n",
    "# Can we find significance?\n",
    "# How many pairs can we afford to test?\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# Below uses the 70 minimum mutated patients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "KeyError",
     "evalue": "0",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-df47e4acb38c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     20\u001b[0m \u001b[0mnumGenes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnumCases\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgenes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpatients\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgeneToCases\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpatientToGenes\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload_mutation_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmutationmatrix\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpatientFile\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgeneFile\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mminFreq\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     21\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 22\u001b[0;31m \u001b[0mcohort_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclusterToProp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmin_cohort\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_patient_cohorts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpartition_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpatientToGenes\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     23\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[0mc0patients\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcohort_dict\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-2-ebed51165c7c>\u001b[0m in \u001b[0;36mload_patient_cohorts\u001b[0;34m(partitionfile, patientToGenes, add_to_closest, delimiter)\u001b[0m\n\u001b[1;32m   1197\u001b[0m         \u001b[0mclusterToProp\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0moldclusterToProp\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msorted_clusters\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1198\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1199\u001b[0;31m     \u001b[0mmin_cohort\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcohort_dict\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1200\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1201\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyError\u001b[0m: 0"
     ]
    }
   ],
   "source": [
    "import mutex as mex\n",
    "import csv\n",
    "import mutex_triangles as met\n",
    "import chisquared as chi\n",
    "import bingenesbypairs as bgbp\n",
    "import time\n",
    "import os\n",
    "\n",
    "mutationmatrix = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.m2'\n",
    "patientFile = '/Users/jlu96/maf/new/BRCA_wustl/shared_patients.plst'\n",
    "partition_file = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.ppf9'\n",
    "cpairfile = '/Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-cpairs-min_cohort.txt'\n",
    "geneFile = None\n",
    "minFreq = 0\n",
    "compute_mutex = True\n",
    "closer_than_distance = 100000000\n",
    "test_minFreq = 3\n",
    "\n",
    "\n",
    "numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "\n",
    "cohort_dict, clusterToProp, min_cohort = load_patient_cohorts(partition_file, patientToGenes)\n",
    "\n",
    "c0patients = cohort_dict[0]\n",
    "c0cohort_dict = {0: c0patients}\n",
    "\n",
    "c0genes, c0geneToCases, c0patientToGenes = get_cluster_gTC_pTG(geneToCases, patientToGenes, c0patients)\n",
    "\n",
    "print \"number genes in smallest cluster is \", len(c0genes)\n",
    "print \"number above threashold \", len([g for g in c0genes if len(c0geneToCases[g]) >= test_minFreq])\n",
    "print \"number patients is \", len(c0patients)\n",
    "\n",
    "t = time.time()\n",
    "genepairs = bgbp.getgenepairs(c0geneToCases, c0genes, test_minFreq=test_minFreq, closer_than_distance=closer_than_distance)\n",
    "print \"Number of pairs is \", len(genepairs), \" retrieved in time : \", time.time() - t\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "cpairsdict, cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, genepairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(cpairsdict)\n",
    "print \"Getting cooccurrence across the whole distribution\"\n",
    "\n",
    "cpairsdict = chi.add_BinomP_cohorts_all_pairs(cpairsdict, geneToCases, patientToGenes, c0cohort_dict, c0patients)\n",
    "cpairsdict = chi.add_BinomP_all_pairs(cpairsdict, geneToCases, patientToGenes)\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "met.writeanydict(cpairsdict, cpairfile, fieldnames=fieldnames)\n",
    "os.system('say \"finished\"')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEKCAYAAAD0Luk/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIiIaLCEQEdFgCYGIiAb7/5Il7VIX9al5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112647d50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of significant pairs  21\n"
     ]
    }
   ],
   "source": [
    "# Plot the p-value distribution\n",
    "pvalues = np.array([cpairsdict[c]['BinomProbability'] for c in cpairsdict])\n",
    "logp = np.log(pvalues)\n",
    "\n",
    "threshold = 0.05/len(logp)\n",
    "\n",
    "plt.figure()\n",
    "plt.hist(logp, bins=50)\n",
    "plt.title(\"Distribution of log p-values\")\n",
    "plt.axvline(x=np.log(threshold), ymin=0, ymax=1000)\n",
    "plt.show()\n",
    "\n",
    "sig = [pvalue for pvalue in pvalues if pvalue < threshold]\n",
    "print \"Number of significant pairs \", len(sig)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Below uses the bottom 15 %, least mutated patients\n",
    "# It filters to those genes mutated in at least10% of those patients\n",
    "# It uses segmentation\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Not importing the Fortran bipartite_edge_swap bindings in permute.py.\n",
      "Number of new patients is  145\n",
      "number genes in smallest cluster is  1267\n",
      "number of genes above threashold  106\n",
      "number patients is  145\n",
      "Number of pairs to test:  4030\n",
      "Number of pairs is  4030  retrieved in time :  4.74499511719\n",
      "number of pairs is  4030\n",
      "Getting cooccurrence across the whole distribution\n",
      "Writing to file...\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Look at top segments. Same filters.\n",
    "\n",
    "\n",
    "# Let's try limiting the significant co-occurrent pairs and see if we can increase our power\n",
    "# Let's consider the 100 least mutated patients and search for co-occurrence within them\n",
    "\n",
    "import mutex as mex\n",
    "import csv\n",
    "import mutex_triangles as met\n",
    "import chisquared as chi\n",
    "import bingenesbypairs as bgbp\n",
    "import time\n",
    "import os\n",
    "import scipy.stats as stats\n",
    "import partition as par\n",
    "\n",
    "mutationmatrix = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-seg-jl.m2'\n",
    "patientFile = '/Users/jlu96/maf/new/BRCA_wustl/shared_patients.plst'\n",
    "partition_file = '/Users/jlu96/maf/new/BRCA_wustl/BRCA_wustl-cna-jl.ppf9'\n",
    "file_prefix = '/Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-'\n",
    "cpairfile = file_prefix + 'cpairs-min_cohort.txt'\n",
    "triplet_file_prefix = file_prefix + 'triplet-'\n",
    "new_cpairfile = file_prefix + \"-cpairs-min_cohort_filtered.txt\"\n",
    "geneFile = None\n",
    "minFreq = 0\n",
    "compute_mutex = True\n",
    "closer_than_distance = 100000000\n",
    "test_minFreq = 0.1\n",
    "minPercentile = 15\n",
    "cpairPercentile = 1\n",
    "mpairPercentile = 1\n",
    "\n",
    "\n",
    "numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "\n",
    "D = [len(patientToGenes[p]) for p in patientToGenes]\n",
    "minThreshold = stats.scoreatpercentile(D, minPercentile)\n",
    "\n",
    "c0patients = [p for p in patientToGenes if len(patientToGenes[p]) <= minThreshold]\n",
    "print \"Number of new patients is \", len(c0patients)\n",
    "test_minFreq = int( test_minFreq * len(c0patients))\n",
    "\n",
    "c0cohort_dict = {0: c0patients}\n",
    "\n",
    "c0genes, c0geneToCases, c0patientToGenes = par.get_cluster_gTC_pTG(geneToCases, patientToGenes, c0patients)\n",
    "\n",
    "print \"number genes in smallest cluster is \", len(c0genes)\n",
    "print \"number of genes above threashold \", len([g for g in c0genes if len(c0geneToCases[g]) >= test_minFreq])\n",
    "print \"number patients is \", len(c0patients)\n",
    "\n",
    "t = time.time()\n",
    "genepairs = bgbp.getgenepairs(c0geneToCases, c0genes, test_minFreq=test_minFreq, closer_than_distance=closer_than_distance)\n",
    "print \"Number of pairs is \", len(genepairs), \" retrieved in time : \", time.time() - t\n",
    "\n",
    "cpairsdict, cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, genepairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(cpairsdict)\n",
    "print \"Getting cooccurrence across the whole distribution\"\n",
    "\n",
    "cpairsdict = chi.add_BinomP_cohorts_all_pairs(cpairsdict, geneToCases, patientToGenes, c0cohort_dict, c0patients)\n",
    "\n",
    "# cpairsdict = chi.add_BinomP_all_pairs(cpairsdict, geneToCases, patientToGenes)\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "met.writeanydict(cpairsdict, cpairfile, fieldnames=fieldnames)\n",
    "os.system('say \"finished\"')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEKCAYAAADzQPVvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1040ddad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAEKCAYAAAAb7IIBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBEBERKX+PwWibU0vMsSxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10b80e5d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top  1 percent of cooccurring pairs:  0.051906486163\n",
      "Top  1 percent of mutually exclusive pairs :  0.104226799535\n",
      "Now number of pairs to test  94\n"
     ]
    }
   ],
   "source": [
    "cpvalues = np.array([cpairsdict[c]['1CBinomProb0'] for c in cpairsdict])\n",
    "logcp = np.log10(cpvalues)\n",
    "mpvalues = np.array([cpairsdict[c]['1MBinomProb0'] for c in cpairsdict])\n",
    "logmp = np.log10(mpvalues)\n",
    "\n",
    "threshold = 0.05/len(logcp)\n",
    "\n",
    "cthreshold = stats.scoreatpercentile(cpvalues, cpairPercentile)\n",
    "mthreshold = stats.scoreatpercentile(mpvalues, mpairPercentile)\n",
    "print \"Top \", cpairPercentile, \"percent of cooccurring pairs: \", cthreshold\n",
    "print \"Top \", mpairPercentile, \"percent of mutually exclusive pairs : \", mthreshold\n",
    "\n",
    "# Let's get the top 10 percent of pairs\n",
    "\n",
    "goodpairs = [c for c in cpairsdict if (cpairsdict[c]['1CBinomProb0'] <= cthreshold or cpairsdict[c]['1MBinomProb0'] <= mthreshold)]\n",
    "print \"Now number of pairs to test \", len(goodpairs)\n",
    "\n",
    "\n",
    "plt.figure()\n",
    "plt.hist(logcp, bins=50)\n",
    "plt.vline(x= np.log10(cthreshold), label=\"Top \" + str(cpairPercentile) + \"Cooccurrence Threshold \")\n",
    "plt.title(\"Distribution of Co-occurring log p-values\", fontsize=20)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "threshold = 0.05/len(logmp)\n",
    "\n",
    "plt.figure()\n",
    "plt.hist(logmp, bins=50)\n",
    "plt.vline(x= np.log10(mthreshold), label=\"Top \" + str(mpairPercentile) + \"Mutually Exclusive Threshold \")\n",
    "plt.title(\"Distribution of Mutually Exclusive log p-values\",fontsize=20)\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of pairs is  94\n",
      "Getting cooccurrence across the whole distribution\n",
      "Writing to file...\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_cpairsdict, new_cgenedict = met.cooccurpairs(numCases, geneToCases, patientToGenes, goodpairs, compute_mutex=compute_mutex)\n",
    "\n",
    "print \"number of pairs is \", len(new_cpairsdict)\n",
    "print \"Getting cooccurrence across the whole distribution\"\n",
    "\n",
    "new_cpairsdict = chi.add_BinomP_cohorts_all_pairs(new_cpairsdict, geneToCases, patientToGenes, c0cohort_dict, c0patients)\n",
    "new_cpairsdict = chi.add_BinomP_all_pairs(new_cpairsdict, geneToCases, patientToGenes)\n",
    "print \"Writing to file...\"\n",
    "\n",
    "fieldnames = (new_cpairsdict.values()[0]).keys()\n",
    "fieldnames.remove('Type')\n",
    "fieldnames.remove('MutationFrequencies')\n",
    "fieldnames.remove('MutationFrequencyDifference')\n",
    "fieldnames.remove('MutationFrequencyDifferenceRatio')\n",
    "fieldnames.remove('CooccurrenceRatio')\n",
    "fieldnames.remove('Coverage')\n",
    "fieldnames.remove('SetScore')\n",
    "fieldnames.remove('AverageOverlapPMN')\n",
    "fieldnames.remove('CombinedScore')\n",
    "fieldnames.remove('Concordance')\n",
    "fieldnames.remove('Somatic')\n",
    "fieldnames.remove('RoundedLogPCov')\n",
    "fieldnames.remove('GeneSet')\n",
    "\n",
    "met.writeanydict(new_cpairsdict, new_cpairfile, fieldnames=fieldnames)\n",
    "os.system('say \"finished\"')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAEKCAYAAADdBdT9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhhzeZmYZcnibmWXI4W1mlqH/B39yqrN/uFqXAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109c22650>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of significant pairs  31\n"
     ]
    }
   ],
   "source": [
    "# Plot the p-value distribution\n",
    "pvalues = np.array([new_cpairsdict[c]['BinomProbability'] for c in new_cpairsdict])\n",
    "logp = np.log(pvalues)\n",
    "\n",
    "threshold = 0.05/len(logp)\n",
    "\n",
    "plt.figure()\n",
    "plt.hist(logp, bins=50)\n",
    "plt.title(\"Distribution of log p-values\")\n",
    "plt.axvline(x=np.log(threshold), ymin=0, ymax=1000)\n",
    "plt.show()\n",
    "\n",
    "sig = [pvalue for pvalue in pvalues if pvalue < threshold]\n",
    "print \"Number of significant pairs \", len(sig)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# add the segment infos\n",
    "\n",
    "bgbp.write_segment_infos(c0genes, \"/Users/jlu96/maf/new/BRCA_wustl/segment_info.txt\")\n",
    "\n",
    "for pair in new_cpairsdict:\n",
    "    info0 = bgbp.get_segment_gene_info(new_cpairsdict[pair]['Gene0'])\n",
    "    new_cpairsdict[pair]['Gene0Loc'] = str(info0['Chromosome']) + ':' + str(info0['Start'])\n",
    "    info1 = bgbp.get_segment_gene_info(new_cpairsdict[pair]['Gene1'])\n",
    "    new_cpairsdict[pair]['Gene1Loc'] = str(info1['Chromosome']) + ':' + str(info1['Start'])\n",
    "    \n",
    "fieldnames += ['Gene0Loc', 'Gene1Loc']\n",
    "met.writeanydict(new_cpairsdict, new_cpairfile, fieldnames=fieldnames)\n",
    "os.system('say \"finished\"')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Results of segmentation. Min cluster size 145. \n",
    "# Tested: 94 pairs. Found co-occurrence\n",
    "\n",
    "\n",
    "31 significant pairs, below:\n",
    "Gene0\tGene1\tMutationFrequency0\tMutationFrequency1\tBinomProbability\n",
    "FAM230Closs\tPEX26_TUBA8_USP18_GGTLC3_PI4KAP1_RN7SKP131_RIMBP3_GGT3P_DGCR6_PRODH_DGCR5_DGCR2_DGCR14_TSSK2_GSC2_SLC25A1_CLTCL1_HIRA_C22orf39_RN7SL168P_MRPL40_UFD1L_CDC45_CLDN5_GP1BB_TBX1_GNB1L_C22orf29_TXNRD2_COMT_MIR4761_ARVCF_TANGO2_MIR185_DGCR8_MIR3618_MIR1306_TRMT2A_RANBP1_ZDHHC8_RTN4R_MIR1286_DGCR6L_FAM230A_USP41_ZNF74_SCARF2_KLHL22_RN7SL812P_MED15_SMPD4P1_POM121L4P_TMEM191A_PI4KA_SERPIND1_SNAP29_CRKL_AIFM3_LZTR1_THAP7_TUBA3FP_P2RX6_SLC7A4_MIR649_P2RX6P_BCRP2_POM121L7_FAM230B_GGT2_RIMBP3B_RN7SKP63_HIC2_TMEM191C_PI4KAP2_RN7SKP221_RIMBP3C_UBE2L3_YDJC_CCDC116_SDF2L1_PPIL2_MIR301B_MIR130B_YPEL1_RN7SL280Ploss\t410\t413\t5.27E-67\n",
    "FAM230Closs\tMAPK1_RNA5SP493loss\t410\t405\t3.20E-64\n",
    "FAM230Closs\tPPM1F_TOP3B_VPREB1_ZNF280B_ZNF280A_PRAME_POM121L1P_GGTLC2_IGLL5_MIR650_IGLJ1_IGLC1_IGLJ2_IGLC2_IGLJ3_IGLC3_IGLJ4_IGLJ5_IGLJ6_IGLJ7_IGLC7_GNAZ_RAB36_BCR_RN7SL263P_CES5AP1_ZDHHC8P1loss\t410\t411\t5.54E-63\n",
    "FAM230Closs\tOR11H1_POTEH_KCNMB3P1_CCT8L2_TPTEP1_XKR3_HSFY1P1_GAB4_CECR7_IL17RA_CECR6_CECR5_CECR1_CECR3_CECR9_RN7SL843P_CECR2_SLC25A18_ATP6V1E1_BCL2L13_BID_LINC00528_MICAL3_MIR648loss\t410\t413\t2.45E-60\n",
    "FAM230Closs\tIGLL1_GUSBP11_RGL4_ZNF70_VPREB3_C22orf15_CHCHD10_MMP11_SMARCB1_DERL3_SLC2A11_RN7SL268P_MIF_GSTT2B_DDTL_DDT_GSTT2_CABIN1_SUSD2_GGT5_POM121L9P_SPECC1L_ADORA2A_UPB1_GUCD1_SNRPD3_GGT1_PIWIL3_SGSM1_TMEM211_KIAA1671_CRYBB3_CRYBB2_LRP5L_CRYBB2P1_ADRBK2_RNA5SP494_MYO18B_RN7SKP169loss\t410\t413\t2.51E-59\n",
    "FAM230Closs\tSEZ6L_RNA5SP495_ASPHD2_HPS4_SRRD_TFIP11_TPST2_MIR548J_CRYBB1_CRYBA4_MIATloss\t410\t424\t4.80E-54\n",
    "FAM230Closs\tTTC28loss\t410\t419\t1.38E-51\n",
    "FAM230Closs\tRN7SL757Ploss\t410\t419\t1.38E-51\n",
    "FAM230Closs\tRN7SL162P_CHEK2_HSCB_CCDC117_XBP1loss\t410\t419\t5.32E-51\n",
    "FAM230Closs\tMN1_PITPNBloss\t410\t419\t5.32E-51\n",
    "FAM230Closs\tZNRF3_C22orf31_KREMEN1_EMID1_RHBDD3_EWSR1_GAS2L1_RASL10A_AP1B1_SNORD125_RFPL1S_RFPL1_NEFH_THOC5_NIPSNAP1_NF2_CABP7_ZMAT5_UQCR10_ASCC2_MTMR3_HORMAD2_LIF_OSM_GATSL3_TBC1D10A_SF3A1_CCDC157_RNF215_SEC14L2_MTFP1_SEC14L3_SEC14L4_SEC14L6_GAL3ST1_PES1_TCN2_SLC35E4_DUSP18_OSBP2_MORC2_TUG1_RN7SL633P_SMTN_INPP5J_PLA2G3_MIR3928_RNF185_LIMK2_PIK3IP1_RNA5SP496_PATZ1_DRG1_EIF4ENIF1_SFI1_PISD_PRR14L_DEPDC5_RN7SL20P_C22orf24_YWHAH_RN7SL305P_SLC5A1_AP1B1P1_C22orf42_RFPL2_SLC5A4_RFPL3_RFPL3S_RTCB_BPIFC_FBXO7_SYN3_RNA5SP497_TIMP3loss\t410\t432\t8.72E-49\n",
    "FAM230Closs\tMIR4764loss\t410\t427\t3.70E-48\n",
    "FAM230Closs\tISXloss\t410\t433\t1.34E-47\n",
    "FAM230Closs\tLARGEloss\t410\t429\t2.52E-47\n",
    "CRK_MYO1C_INPP5K_PITPNA_SLC43A2_RN7SL105P_SCARF1_RILP_PRPF8_TLCD2_MIR22HG_WDR81_SERPINF2_SERPINF1_SMYD4_RPA1_RTN4RL1_DPH1_OVCA2_MIR132_MIR212_HIC1_SMG6_RN7SL624P_SRR_TSR1_SNORD91B_SNORD91A_SGSM2_MNT_METTL16_RN7SL33P_PAFAH1B1_RN7SL608P_CLUH_MIR1253_RAP1GAP2_OR1D5_OR1D2_OR1G1_OR1A2_OR1A1_OR1D4_OR3A2_OR3A1_OR3A4P_OR1E1_OR3A3_OR1E2_SPATA22_ASPA_TRPV3_TRPV1_SHPK_CTNS_TAX1BP3_EMC6_P2RX5_ITGAE_GSG2_C17orf85_CAMKK1_P2RX1_ATP2A3_ZZEF1_RNA5SP434_CYB5D2_ANKFY1_UBE2G1_RN7SL774P_SPNS3_SPNS2_MYBBP1A_GGT6_SMTNL2_ALOX15_PELP1_ARRB2_MED11_CXCL16_ZMYND15_TM4SF5_VMO1_GLTPD2_PSMB6_PLD2_MINK1_RN7SL784P_CHRNE_C17orf107_GP1BA_SLC25A11_RNF167_PFN1_ENO3_SPAG7_CAMTA2_INCA1_KIF1C_SLC52A1_ZFP3_ZNF232_USP6_ZNF594_SCIMP_RABEP1_NUP88_RPAIN_C1QBP_DHX33_DERL2_MIS12_NLRP1loss\tRN7SL605Ploss\t580\t564\t1.94E-46\n",
    "DOC2B_RPH3AL_C17orf97_FAM101B_VPS53_FAM57A_GEMIN4_DBIL5P_GLOD4_RNMTL1_NXN_TIMM22_ABR_MIR3183_BHLHA9_TUSC5loss\tRN7SL605Ploss\t565\t564\t3.87E-46\n",
    "YWHAEloss\tRN7SL605Ploss\t566\t564\t7.08E-46\n",
    "FAM230Closs\tHMGXB4_TOM1_MIR3909_HMOX1_MCM5_RASD2_MB_APOL6_APOL5_RBFOX2_APOL3_APOL4_APOL2_APOL1_MYH9loss\t410\t448\t3.19E-45\n",
    "WSCD1loss\tRN7SL605Ploss\t574\t564\t9.93E-43\n",
    "FAM230Closs\tLINC00898loss\t410\t442\t3.32E-41\n",
    "FAM230Closs\tTBC1D22Aloss\t410\t449\t1.90E-40\n",
    "FAM230Closs\tMIR3201_FAM19A5_MIR4535_C22orf34_MIR3667_RN7SKP252_BRD1_ZBED4_ALG12_CRELD2_PIM3_IL17REL_TTLL8_MLC1_MOV10L1_PANX2_TRABD_TUBGCP6_HDAC10_MAPK12_MAPK11_PLXNB2_DENND6B_PPP6R2_RN7SL500P_SBF1_ADM2_MIOX_LMF2_NCAPH2_SCO2_TYMP_ODF3B_KLHDC7B_SYCE3_CPT1B_CHKB_MAPK8IP2_ARSA_SHANK3_ACR_RABL2Bloss\t410\t438\t3.66E-40\n",
    "AIPL1_FAM64A_KIAA0753_RNA5SP435_TXNDC17_MED31_C17orf100_SLC13A5_XAF1_FBXO39_TEKT1_ALOX12P2_ALOX12_RNASEK_C17orf49_MIR497HG_BCL6B_SLC16A13_SLC16A11_CLEC10A_ASGR2_ASGR1_DLG4_ACADVL_MIR324_DVL2_PHF23_GABARAP_CTDNEP1_ELP5_CLDN7_SLC2A4_YBX2_EIF5A_GPS2_NEURL4_ACAP1_KCTD11_TMEM95_TNK1_TMEM256_NLGN2_SPEM1_C17orf74_TMEM102_FGF11_CHRNB1_ZBTB4_SLC35G6_POLR2A_TNFSF12_TNFSF13_SENP3_EIF4A1_SNORD10_CD68_MPDU1_SOX15_FXR2_SHBG_SAT2_ATP1B2_TP53_WRAP53_EFNB3_DNAH2_RPL29P2_KDM6B_TMEM88loss\tRN7SL605Ploss\t596\t564\t9.45E-40\n",
    "CYB5D1_CHD3_KCNAB3_TRAPPC1_CNTROB_GUCY2D_ALOX15B_ALOX12B_MIR4314_ALOXE3_HES7_PER1_VAMP2_TMEM107_SNORD118_C17orf59_AURKB_LINC00324_CTC1_PFAS_SLC25A35_RANGRF_ARHGEF15_ODF4_KRBA2_RPL26_RNF222_NDEL1_MYH10_CCDC42_SPDYE4_MFSD6L_PIK3R6_PIK3R5_NTN1loss\tRN7SL605Ploss\t591\t564\t1.05E-39\n",
    "DNAH9_ZNF18_RPL21P122loss\tRN7SL605Ploss\t583\t564\t4.32E-39\n",
    "MAP2K4_MIR744loss\tRN7SL605Ploss\t583\t564\t4.32E-39\n",
    "STX8_USP43_DHRS7C_GLP2R_RCVRN_GAS7loss\tRN7SL605Ploss\t591\t564\t1.12E-38\n",
    "FAM230Closs\tTXN2_FOXRED2_EIF3D_CACNG2_IFT27_PVALB_NCF4_CSF2RB_TEX33_TST_MPST_KCTD17_RN7SKP214_TMPRSS6_IL2RB_C1QTNF6_SSTR3_RAC2_CYTH4_ELFN2_MFNG_CARD10_CDC42EP1_LGALS2_GGA1_SH3BP1_PDXP_RN7SL385P_LGALS1_NOL12_TRIOBP_H1F0_GCAT_GALR3_ANKRD54_MIR658_MIR659_EIF3L_MICALL1_C22orf23_POLR2F_SOX10_MIR4534_PICK1_SLC16A8_BAIAP2L2_PLA2G6_MAFF_TMEM184B_RN7SL704P_CSNK1E_KCNJ4_KDELR3_DDX17_DMC1_FAM227A_CBY1_TOMM22_JOSD1_GTPBP1_SUN2_DNAL4_NPTXR_CBX6_APOBEC3A_APOBEC3B_APOBEC3C_APOBEC3D_APOBEC3F_APOBEC3G_APOBEC3H_CBX7_PDGFB_RPL3_SNORD83B_SNORD83A_SYNGR1_TAB1_MGAT3_ATF4_RPS19BP1_CACNA1I_ENTHD1_RN7SKP210_GRAP2_FAM83F_TNRC6B_ADSL_SGSM3_MKL1_MCHR1_SLC25A17_MIR4766_ST13_XPNPEP3_DNAJB7_RBX1_EP300_MIR1281_L3MBTL2_CHADL_RANGAP1_ZC3H7B_TEF_TOB2_PHF5A_ACO2_POLR3H_CSDC2_PMM1_DESI1_XRCC6_NHP2L1_C22orf46_MEI1_RNU6ATAC22P_CCDC134_SREBF2_MIR33A_SHISA8_TNFRSF13C_MIR378I_CENPM_LINC00634_WBP2NL_NAGA_FAM109B_SMDT1_NDUFA6_CYP2D6_TCF20_NFAM1_SERHL_RRP7A_SERHL2_RN7SKP80_POLDIP3_RNU12_CYB5R3_ATP5L2_A4GALT_ARFGAP3_PACSIN2_TTLL1_BIK_MCAT_TSPO_TTLL12_SCUBE1_MPPED1_EFCAB6_SULT4A1_PNPLA5_PNPLA3_SAMM50_PARVB_PARVG_KIAA1644_LDOC1L_LINC00207_LINC00229_PRR5_ARHGAP8_PHF21B_NUP50_KIAA0930_MIR1249_UPK3A_FAM118A_SMC1B_RIBC2_FBLN1_ATXN10_MIR4762_WNT7B_LINC00899_MIR3619_MIRLET7A3_MIR4763_MIRLET7B_PPARA_CDPF1_PKDREJ_TTC38_GTSE1_TRMU_CELSR1_GRAMD4_CERKloss\t410\t476\t1.13E-38\n",
    "MYH13_MYH8_MYH4_MYH1_MYH2_MYH3_SCO1_ADPRM_TMEM220_LINC00675_PIRT_RN7SL601P_SHISA6loss\tRN7SL605Ploss\t585\t564\t1.25E-38\n",
    "LINC00670_MYOCD_ARHGAP44_MIR1269B_ELAC2loss\tRN7SL605Ploss\t579\t564\t1.96E-38\n",
    "HS3ST3A1_MIR548H3_COX10_CDRT15_HS3ST3B1_CDRT7_CDRT8_PMP22_MIR4731_TEKT3_RN7SL792P_CDRT4_TVP23C_CDRT1_TRIM16_ZNF286A_TBC1D26_RNA5SP436_ADORA2B_ZSWIM7_TTC19_NCOR1_RN7SL442P_PIGL_MIR1288_CENPV_UBB_TRPV2_ZNF287_ZNF624_CCDC144A_RN7SL620P_USP32P1_KRT16P2_KRT17P1_TBC1D27_TNFRSF13Bloss\tRN7SL605Ploss\t575\t564\t6.06E-38\n",
    "\n",
    "Locations:\n",
    "Gene0Loc\tGene1Loc\n",
    "13:18195297\t22:18077920\n",
    "13:18195297\t22:21754500\n",
    "13:18195297\t22:21919420\n",
    "13:18195297\t22:15528158\n",
    "13:18195297\t22:23573125\n",
    "13:18195297\t22:26169474\n",
    "13:18195297\t22:27978014\n",
    "13:18195297\t22:28056153\n",
    "13:18195297\t22:28642979\n",
    "13:18195297\t22:27748277\n",
    "13:18195297\t22:28883592\n",
    "13:18195297\t22:33436582\n",
    "13:18195297\t22:35066136\n",
    "13:18195297\t22:33162226\n",
    "17:1420689\t11:57528085\n",
    "17:142789\t11:57528085\n",
    "17:1344272\t11:57528085\n",
    "13:18195297\t22:35257452\n",
    "17:5772234\t11:57528085\n",
    "13:18195297\t22:47621043\n",
    "13:18195297\t22:46762617\n",
    "13:18195297\t22:48274364\n",
    "17:6393693\t11:57528085\n",
    "17:7857746\t11:57528085\n",
    "17:11598431\t11:57528085\n",
    "17:12020824\t11:57528085\n",
    "17:9250471\t11:57528085\n",
    "13:18195297\t22:36467036\n",
    "17:10298084\t11:57528085\n",
    "17:12549782\t11:57528085\n",
    "17:13495689\t11:57528085\n",
    "\n",
    "This looks good.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of pairs  94\n",
      "Pair info of triplets initialized\n",
      "36  triplets calculated in  0.0374140739441\n",
      "Including sorting time  0.037712097168\n",
      "19  of type  CooccurringMutuallyExclusiveMutuallyExclusive\n",
      "Triplets written to  /Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-triplet-CooccurringMutuallyExclusiveMutuallyExclusive.txt_Triplets.tsv\n",
      "7  of type  CooccurringCooccurringMutuallyExclusive\n",
      "Triplets written to  /Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-triplet-CooccurringCooccurringMutuallyExclusive.txt_Triplets.tsv\n",
      "10  of type  CooccurringCooccurringCooccurring\n",
      "Triplets written to  /Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-triplet-CooccurringCooccurringCooccurring.txt_Triplets.tsv\n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.433368310598 p_g_m 0.383001049318\n",
      "p_cmm  0.238689552735 expected  227.471143757\n",
      "observed prop 0.495278069255 num overlap  472 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 p_g_m 0.383001049318\n",
      "p_cmm  0.483671709926 expected  460.939139559\n",
      "observed prop 0.495278069255 num overlap  472 out of  953 total \n",
      "Coccur pair :  410 427 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.448058761805 p_g_m 0.383001049318\n",
      "p_cmm  0.239383225667 expected  228.132214061\n",
      "observed prop 0.460650577125 num overlap  439 out of  953 total \n",
      "Coccur pair :  410 427 Mutex:  365\n",
      "p_g_cs 0.397691500525 p_g_m 0.383001049318\n",
      "p_cmm  0.476060025831 expected  453.685204617\n",
      "observed prop 0.460650577125 num overlap  439 out of  953 total \n",
      "Coccur pair :  410 429 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.450157397692 p_g_m 0.383001049318\n",
      "p_cmm  0.2394823218 expected  228.226652676\n",
      "observed prop 0.456453305352 num overlap  435 out of  953 total \n",
      "Coccur pair :  410 429 Mutex:  365\n",
      "p_g_cs 0.396642182581 p_g_m 0.383001049318\n",
      "p_cmm  0.475814487634 expected  453.451206716\n",
      "observed prop 0.456453305352 num overlap  435 out of  953 total \n",
      "Coccur pair :  410 442 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.463798530955 p_g_m 0.383001049318\n",
      "p_cmm  0.240126446666 expected  228.840503673\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 442 Mutex:  365\n",
      "p_g_cs 0.389296956978 p_g_m 0.383001049318\n",
      "p_cmm  0.474095720258 expected  451.813221406\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.400839454355 p_g_m 0.383001049318\n",
      "p_cmm  0.476796640421 expected  454.387198321\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.433368310598 p_g_m 0.383001049318\n",
      "p_cmm  0.238689552735 expected  227.471143757\n",
      "observed prop 0.482686253935 num overlap  460 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.414480587618 p_g_m 0.383001049318\n",
      "p_cmm  0.479988636977 expected  457.429171039\n",
      "observed prop 0.482686253935 num overlap  460 out of  953 total \n",
      "Coccur pair :  410 411 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.431269674711 p_g_m 0.383001049318\n",
      "p_cmm  0.238590456602 expected  227.376705142\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "Coccur pair :  410 411 Mutex:  365\n",
      "p_g_cs 0.420776495278 p_g_m 0.383001049318\n",
      "p_cmm  0.481461866156 expected  458.833158447\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "Coccur pair :  410 476 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.499475341028 p_g_m 0.383001049318\n",
      "p_cmm  0.24181108093 expected  230.445960126\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 476 Mutex:  365\n",
      "p_g_cs 0.401888772298 p_g_m 0.383001049318\n",
      "p_cmm  0.477042178617 expected  454.621196222\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.465897166842 num overlap  444 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.399790136411 p_g_m 0.383001049318\n",
      "p_cmm  0.476551102224 expected  454.15320042\n",
      "observed prop 0.465897166842 num overlap  444 out of  953 total \n",
      "Coccur pair :  410 449 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.471143756558 p_g_m 0.383001049318\n",
      "p_cmm  0.240473283132 expected  229.171038825\n",
      "observed prop 0.451206715635 num overlap  430 out of  953 total \n",
      "Coccur pair :  410 449 Mutex:  365\n",
      "p_g_cs 0.391395592865 p_g_m 0.383001049318\n",
      "p_cmm  0.474586796651 expected  452.281217209\n",
      "observed prop 0.451206715635 num overlap  430 out of  953 total \n",
      "Coccur pair :  410 438 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.459601259182 p_g_m 0.383001049318\n",
      "p_cmm  0.2399282544 expected  228.651626443\n",
      "observed prop 0.447009443861 num overlap  426 out of  953 total \n",
      "Coccur pair :  410 438 Mutex:  365\n",
      "p_g_cs 0.384050367261 p_g_m 0.383001049318\n",
      "p_cmm  0.472868029275 expected  450.643231899\n",
      "observed prop 0.447009443861 num overlap  426 out of  953 total \n",
      "Coccur pair :  410 448 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.470094438615 p_g_m 0.383001049318\n",
      "p_cmm  0.240423735065 expected  229.123819517\n",
      "observed prop 0.461699895068 num overlap  440 out of  953 total \n",
      "Coccur pair :  410 448 Mutex:  365\n",
      "p_g_cs 0.402938090241 p_g_m 0.383001049318\n",
      "p_cmm  0.477287716814 expected  454.855194124\n",
      "observed prop 0.461699895068 num overlap  440 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.464847848898 num overlap  443 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.399790136411 p_g_m 0.383001049318\n",
      "p_cmm  0.476551102224 expected  454.15320042\n",
      "observed prop 0.464847848898 num overlap  443 out of  953 total \n",
      "Coccur pair :  410 433 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.454354669465 p_g_m 0.383001049318\n",
      "p_cmm  0.239680514067 expected  228.415529906\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 433 Mutex:  365\n",
      "p_g_cs 0.399790136411 p_g_m 0.383001049318\n",
      "p_cmm  0.476551102224 expected  454.15320042\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 432 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.453305351522 p_g_m 0.383001049318\n",
      "p_cmm  0.239630966 expected  228.368310598\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 432 Mutex:  365\n",
      "p_g_cs 0.401888772298 p_g_m 0.383001049318\n",
      "p_cmm  0.477042178617 expected  454.621196222\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.400839454355 p_g_m 0.383001049318\n",
      "p_cmm  0.476796640421 expected  454.387198321\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 424 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.444910807975 p_g_m 0.383001049318\n",
      "p_cmm  0.239234581467 expected  227.990556139\n",
      "observed prop 0.472193074502 num overlap  450 out of  953 total \n",
      "Coccur pair :  410 424 Mutex:  365\n",
      "p_g_cs 0.409233997901 p_g_m 0.383001049318\n",
      "p_cmm  0.478760945994 expected  456.259181532\n",
      "observed prop 0.472193074502 num overlap  450 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.433368310598 p_g_m 0.383001049318\n",
      "p_cmm  0.238689552735 expected  227.471143757\n",
      "observed prop 0.480587618048 num overlap  458 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.416579223505 p_g_m 0.383001049318\n",
      "p_cmm  0.48047971337 expected  457.897166842\n",
      "observed prop 0.480587618048 num overlap  458 out of  953 total \n",
      "Coccur pair :  410 405 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.424973767051 p_g_m 0.383001049318\n",
      "p_cmm  0.238293168202 expected  227.093389297\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "Coccur pair :  410 405 Mutex:  365\n",
      "p_g_cs 0.419727177335 p_g_m 0.383001049318\n",
      "p_cmm  0.48121632796 expected  458.599160546\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "[4.6018466838794629e-66, 2.2732617823899843e-50, 1.5792674494309096e-48, 1.2999042568991835e-45, 3.5065305274658016e-53, 3.5427359777070553e-60, 3.1666592115289563e-63, 6.9401758233186083e-45, 1.2282207405392256e-52, 5.0503605965456466e-46, 1.3083400304917592e-44, 3.1069102022201346e-50, 2.4868579803310521e-52, 4.111061589494067e-51, 3.9558686804913671e-51, 3.5065305274658016e-53, 2.3422524970912992e-55, 2.8066765268020981e-59, 2.5958333653137317e-63]\n",
      "[0.23800684121758911, 0.17354352097606404, 0.12149421845662704, 0.072188557275518123, 0.27972758032205808, 0.43564308425428999, 0.32502595491233122, 0.049110062288141014, 0.25837568090533358, 0.076772550650045832, 0.055986633193093729, 0.17352856009159795, 0.23785989414945108, 0.19964389431079288, 0.19072343568438904, 0.27972758032205808, 0.34864363845427337, 0.5, 0.32499101827171334]\n"
     ]
    }
   ],
   "source": [
    "# Search for triplets\n",
    "import mutex_triangles as met\n",
    "import edgereader as edg\n",
    "reload(chi)\n",
    "reload(met)\n",
    "\n",
    "pairsdict = new_cpairsdict.copy()\n",
    "genedict = edg.get_gene_dict(pairsdict)\n",
    "Triplets, pairsdict_Triplets, sorted_pairs, genesdict_Triplets, sorted_genes = met.getTriplets(pairsdict, genedict, pairsdict, numCases, geneToCases=geneToCases, patientToGenes=patientToGenes)\n",
    "Triplet_dict = met.sort_triplets_by_type(Triplets)\n",
    "\n",
    "for t_type in Triplet_dict:\n",
    "    print len(Triplet_dict[t_type]), \" of type \",  t_type\n",
    "    met.writeTriplets(Triplet_dict[t_type], triplet_file_prefix + t_type + '.txt')\n",
    "\n",
    "# Let's built a nonparametric method to test for significance\n",
    "\n",
    "pvalues = []\n",
    "pvalues_ab = []\n",
    "for i in range(len(Triplets)):\n",
    "    \n",
    "    t_genes = Triplets[i].genes\n",
    "    if Triplets[i].type == 'CooccurringMutuallyExclusiveMutuallyExclusive':\n",
    "        pvalues.append(chi.get_triplet_BinomP(t_genes, geneToCases, patientToGenes,\n",
    "                                     cpairs=Triplets[i].cpairs,\n",
    "                                     mpairs=Triplets[i].mpairs))\n",
    "        pvalues_ab.append(chi.get_triplet_BinomP_ab(t_genes, geneToCases, patientToGenes,\n",
    "                                             cpairs=Triplets[i].cpairs,\n",
    "                                             mpairs=Triplets[i].mpairs))\n",
    "\n",
    "               \n",
    "print pvalues\n",
    "print pvalues_ab\n",
    "\n",
    "\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "with open(triplet_file_prefix + 'CooccurringMutuallyExclusiveMutuallyExclusive' + '_pvalues.txt', 'w') as csvfile:\n",
    "    writer = csv.writer(csvfile,delimiter='\\t')\n",
    "    writer.writerow(['P(A)P(B) pvalue', 'P(AB) pvalue'])\n",
    "    for pvalue, pvalue_ab in zip(pvalues, pvalues_ab):\n",
    "        writer.writerow([pvalue, pvalue_ab])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZEAAAEnCAYAAABosn4/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1176dd390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "logpvalues = np.log10(pvalues)\n",
    "plt.scatter(logpvalues, pvalues_ab)\n",
    "plt.xlabel(\"Log PValues using P(A)P(B)\", fontsize=20)\n",
    "plt.ylabel(\"Pvalues using P(A&B)\", fontsize=20)\n",
    "plt.title(\"Triplet Pvalues\", fontsize=20)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Triplets analysis\n",
    "\n",
    "One of the triplets:\n",
    "CMM: 151 in all, 259 have 2, 213 have only the remaining"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Mutual Exclusivity/Co-occurrence: Statistical model\n",
    "=====================================================\n",
    "\n",
    "Let D represent our data. Let P(A), P(B) be bernoulli random variables, and assume our observations (patients)are all i.i.d.\n",
    "Let X_A be the count of A in the data. Let X_B be the count of B in the data. Both are binomial random variables.\n",
    "Let X_{A&B} be the count of A & B in the data.\n",
    "Let n be the total observations (i.e. patients) we have.\n",
    "\n",
    "\n",
    "Define mutual exclusivity b/n A and B as\n",
    "\n",
    "$$M_{A,B} = P(A&B) < P(A)P(B)$$\n",
    "\n",
    "Define co-occurrence b/n A and B as \n",
    "\n",
    "$$C_{A,B} = P(A&B) > P(A)P(B)$$\n",
    "\n",
    "Define triplet co-occurrence b/n A, B, and C as\n",
    "\n",
    "$$CCC_{A, B, C} = P(A&B&C) > P(A)P(B)P(C)$$\n",
    "\n",
    "Define CCM b/n A, B, and C (A&B are mutex) as\n",
    "\n",
    "$$CCM_{A, B, C} = P(A&!B&C) + P(!A&B&C) > P(A)P(!B)P(C) + P(!A)P(B)P(C)$$\n",
    "\n",
    "Define CMM b/n A, B, and C (A&B cooccur) as\n",
    "\n",
    "$$CMM_{A, B, C} = P(A&B&!C) + P(!A&!B&C) > P(A)P(B)P(!C) + P(!A)P(!B)P(C)$$\n",
    "\n",
    "Define triplet mutual exclusviity b/n A, B, and C as:\n",
    "\n",
    "$$MMM_{A, B, C} = P(A&!B&!C) + P(!A&B&!C) + P(!A&!B&C) > P(A)P(!B)P(!C) + P(!A)P(B)P(!C) + P(!A)P(!B)P(C)$$.\n",
    "\n",
    "\n",
    "\n",
    "### Frequentist formulation\n",
    "We test for M_{A,B} under $H_0: P(A&B) = P(A)P(B)$.\n",
    "H_0 assumes: A & B are independent Bernoulli random variables\n",
    "\n",
    "We estimate P(A), and P(B) as p(A)=X_A/n, and p(B)=X_B/n, the data.\n",
    "We then calculate $P(X_{A&B} | P(A&B) = P(A)P(B), P(A) = p(A), P(B) = p(B))$. If this is below some threshold, we accept.a\n",
    "\n",
    "\n",
    "### Bayesian formulation\n",
    "Using the same estimates as above,\n",
    "we then calculate $P(X_{A&B} | P(A&B) = P(A)P(B), P(A) = p(A), P(B) = p(B))$.\n",
    "We can also calculate $P(X_{A&B} | P(A&B) > P(A)P(B), P(A) = p(A), P(B) = p(B))$. This is conditioning on the event of co_occurrence\n",
    "\n",
    "\n",
    "We can then compute Bayes Factors: \n",
    "$P(X_{A&B} | P(A&B) > p(A)p(B))/P(X_{A&B} | P(A&B) = p(A)p(B))$\n",
    "\n",
    "\n",
    "Should we integrate out over events P(A) != p(A), P(B) != p(B) to get:\n",
    "P(X_{A&B} | P(A&B) = P(A)P(B))?\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "We have co-occurrence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# BRCA wustl: rerun"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Results of segmentation. Min cluster size 145. \n",
    "# Tested: 94 pairs. Found co-occurrence\n",
    "\n",
    "\n",
    "31 significant pairs, below:\n",
    "Gene0\tGene1\tMutationFrequency0\tMutationFrequency1\tBinomProbability\n",
    "FAM230Closs\tPEX26_TUBA8_USP18_GGTLC3_PI4KAP1_RN7SKP131_RIMBP3_GGT3P_DGCR6_PRODH_DGCR5_DGCR2_DGCR14_TSSK2_GSC2_SLC25A1_CLTCL1_HIRA_C22orf39_RN7SL168P_MRPL40_UFD1L_CDC45_CLDN5_GP1BB_TBX1_GNB1L_C22orf29_TXNRD2_COMT_MIR4761_ARVCF_TANGO2_MIR185_DGCR8_MIR3618_MIR1306_TRMT2A_RANBP1_ZDHHC8_RTN4R_MIR1286_DGCR6L_FAM230A_USP41_ZNF74_SCARF2_KLHL22_RN7SL812P_MED15_SMPD4P1_POM121L4P_TMEM191A_PI4KA_SERPIND1_SNAP29_CRKL_AIFM3_LZTR1_THAP7_TUBA3FP_P2RX6_SLC7A4_MIR649_P2RX6P_BCRP2_POM121L7_FAM230B_GGT2_RIMBP3B_RN7SKP63_HIC2_TMEM191C_PI4KAP2_RN7SKP221_RIMBP3C_UBE2L3_YDJC_CCDC116_SDF2L1_PPIL2_MIR301B_MIR130B_YPEL1_RN7SL280Ploss\t410\t413\t5.27E-67\n",
    "FAM230Closs\tMAPK1_RNA5SP493loss\t410\t405\t3.20E-64\n",
    "FAM230Closs\tPPM1F_TOP3B_VPREB1_ZNF280B_ZNF280A_PRAME_POM121L1P_GGTLC2_IGLL5_MIR650_IGLJ1_IGLC1_IGLJ2_IGLC2_IGLJ3_IGLC3_IGLJ4_IGLJ5_IGLJ6_IGLJ7_IGLC7_GNAZ_RAB36_BCR_RN7SL263P_CES5AP1_ZDHHC8P1loss\t410\t411\t5.54E-63\n",
    "FAM230Closs\tOR11H1_POTEH_KCNMB3P1_CCT8L2_TPTEP1_XKR3_HSFY1P1_GAB4_CECR7_IL17RA_CECR6_CECR5_CECR1_CECR3_CECR9_RN7SL843P_CECR2_SLC25A18_ATP6V1E1_BCL2L13_BID_LINC00528_MICAL3_MIR648loss\t410\t413\t2.45E-60\n",
    "FAM230Closs\tIGLL1_GUSBP11_RGL4_ZNF70_VPREB3_C22orf15_CHCHD10_MMP11_SMARCB1_DERL3_SLC2A11_RN7SL268P_MIF_GSTT2B_DDTL_DDT_GSTT2_CABIN1_SUSD2_GGT5_POM121L9P_SPECC1L_ADORA2A_UPB1_GUCD1_SNRPD3_GGT1_PIWIL3_SGSM1_TMEM211_KIAA1671_CRYBB3_CRYBB2_LRP5L_CRYBB2P1_ADRBK2_RNA5SP494_MYO18B_RN7SKP169loss\t410\t413\t2.51E-59\n",
    "FAM230Closs\tSEZ6L_RNA5SP495_ASPHD2_HPS4_SRRD_TFIP11_TPST2_MIR548J_CRYBB1_CRYBA4_MIATloss\t410\t424\t4.80E-54\n",
    "FAM230Closs\tTTC28loss\t410\t419\t1.38E-51\n",
    "FAM230Closs\tRN7SL757Ploss\t410\t419\t1.38E-51\n",
    "FAM230Closs\tRN7SL162P_CHEK2_HSCB_CCDC117_XBP1loss\t410\t419\t5.32E-51\n",
    "FAM230Closs\tMN1_PITPNBloss\t410\t419\t5.32E-51\n",
    "FAM230Closs\tZNRF3_C22orf31_KREMEN1_EMID1_RHBDD3_EWSR1_GAS2L1_RASL10A_AP1B1_SNORD125_RFPL1S_RFPL1_NEFH_THOC5_NIPSNAP1_NF2_CABP7_ZMAT5_UQCR10_ASCC2_MTMR3_HORMAD2_LIF_OSM_GATSL3_TBC1D10A_SF3A1_CCDC157_RNF215_SEC14L2_MTFP1_SEC14L3_SEC14L4_SEC14L6_GAL3ST1_PES1_TCN2_SLC35E4_DUSP18_OSBP2_MORC2_TUG1_RN7SL633P_SMTN_INPP5J_PLA2G3_MIR3928_RNF185_LIMK2_PIK3IP1_RNA5SP496_PATZ1_DRG1_EIF4ENIF1_SFI1_PISD_PRR14L_DEPDC5_RN7SL20P_C22orf24_YWHAH_RN7SL305P_SLC5A1_AP1B1P1_C22orf42_RFPL2_SLC5A4_RFPL3_RFPL3S_RTCB_BPIFC_FBXO7_SYN3_RNA5SP497_TIMP3loss\t410\t432\t8.72E-49\n",
    "FAM230Closs\tMIR4764loss\t410\t427\t3.70E-48\n",
    "FAM230Closs\tISXloss\t410\t433\t1.34E-47\n",
    "FAM230Closs\tLARGEloss\t410\t429\t2.52E-47\n",
    "CRK_MYO1C_INPP5K_PITPNA_SLC43A2_RN7SL105P_SCARF1_RILP_PRPF8_TLCD2_MIR22HG_WDR81_SERPINF2_SERPINF1_SMYD4_RPA1_RTN4RL1_DPH1_OVCA2_MIR132_MIR212_HIC1_SMG6_RN7SL624P_SRR_TSR1_SNORD91B_SNORD91A_SGSM2_MNT_METTL16_RN7SL33P_PAFAH1B1_RN7SL608P_CLUH_MIR1253_RAP1GAP2_OR1D5_OR1D2_OR1G1_OR1A2_OR1A1_OR1D4_OR3A2_OR3A1_OR3A4P_OR1E1_OR3A3_OR1E2_SPATA22_ASPA_TRPV3_TRPV1_SHPK_CTNS_TAX1BP3_EMC6_P2RX5_ITGAE_GSG2_C17orf85_CAMKK1_P2RX1_ATP2A3_ZZEF1_RNA5SP434_CYB5D2_ANKFY1_UBE2G1_RN7SL774P_SPNS3_SPNS2_MYBBP1A_GGT6_SMTNL2_ALOX15_PELP1_ARRB2_MED11_CXCL16_ZMYND15_TM4SF5_VMO1_GLTPD2_PSMB6_PLD2_MINK1_RN7SL784P_CHRNE_C17orf107_GP1BA_SLC25A11_RNF167_PFN1_ENO3_SPAG7_CAMTA2_INCA1_KIF1C_SLC52A1_ZFP3_ZNF232_USP6_ZNF594_SCIMP_RABEP1_NUP88_RPAIN_C1QBP_DHX33_DERL2_MIS12_NLRP1loss\tRN7SL605Ploss\t580\t564\t1.94E-46\n",
    "DOC2B_RPH3AL_C17orf97_FAM101B_VPS53_FAM57A_GEMIN4_DBIL5P_GLOD4_RNMTL1_NXN_TIMM22_ABR_MIR3183_BHLHA9_TUSC5loss\tRN7SL605Ploss\t565\t564\t3.87E-46\n",
    "YWHAEloss\tRN7SL605Ploss\t566\t564\t7.08E-46\n",
    "FAM230Closs\tHMGXB4_TOM1_MIR3909_HMOX1_MCM5_RASD2_MB_APOL6_APOL5_RBFOX2_APOL3_APOL4_APOL2_APOL1_MYH9loss\t410\t448\t3.19E-45\n",
    "WSCD1loss\tRN7SL605Ploss\t574\t564\t9.93E-43\n",
    "FAM230Closs\tLINC00898loss\t410\t442\t3.32E-41\n",
    "FAM230Closs\tTBC1D22Aloss\t410\t449\t1.90E-40\n",
    "FAM230Closs\tMIR3201_FAM19A5_MIR4535_C22orf34_MIR3667_RN7SKP252_BRD1_ZBED4_ALG12_CRELD2_PIM3_IL17REL_TTLL8_MLC1_MOV10L1_PANX2_TRABD_TUBGCP6_HDAC10_MAPK12_MAPK11_PLXNB2_DENND6B_PPP6R2_RN7SL500P_SBF1_ADM2_MIOX_LMF2_NCAPH2_SCO2_TYMP_ODF3B_KLHDC7B_SYCE3_CPT1B_CHKB_MAPK8IP2_ARSA_SHANK3_ACR_RABL2Bloss\t410\t438\t3.66E-40\n",
    "AIPL1_FAM64A_KIAA0753_RNA5SP435_TXNDC17_MED31_C17orf100_SLC13A5_XAF1_FBXO39_TEKT1_ALOX12P2_ALOX12_RNASEK_C17orf49_MIR497HG_BCL6B_SLC16A13_SLC16A11_CLEC10A_ASGR2_ASGR1_DLG4_ACADVL_MIR324_DVL2_PHF23_GABARAP_CTDNEP1_ELP5_CLDN7_SLC2A4_YBX2_EIF5A_GPS2_NEURL4_ACAP1_KCTD11_TMEM95_TNK1_TMEM256_NLGN2_SPEM1_C17orf74_TMEM102_FGF11_CHRNB1_ZBTB4_SLC35G6_POLR2A_TNFSF12_TNFSF13_SENP3_EIF4A1_SNORD10_CD68_MPDU1_SOX15_FXR2_SHBG_SAT2_ATP1B2_TP53_WRAP53_EFNB3_DNAH2_RPL29P2_KDM6B_TMEM88loss\tRN7SL605Ploss\t596\t564\t9.45E-40\n",
    "CYB5D1_CHD3_KCNAB3_TRAPPC1_CNTROB_GUCY2D_ALOX15B_ALOX12B_MIR4314_ALOXE3_HES7_PER1_VAMP2_TMEM107_SNORD118_C17orf59_AURKB_LINC00324_CTC1_PFAS_SLC25A35_RANGRF_ARHGEF15_ODF4_KRBA2_RPL26_RNF222_NDEL1_MYH10_CCDC42_SPDYE4_MFSD6L_PIK3R6_PIK3R5_NTN1loss\tRN7SL605Ploss\t591\t564\t1.05E-39\n",
    "DNAH9_ZNF18_RPL21P122loss\tRN7SL605Ploss\t583\t564\t4.32E-39\n",
    "MAP2K4_MIR744loss\tRN7SL605Ploss\t583\t564\t4.32E-39\n",
    "STX8_USP43_DHRS7C_GLP2R_RCVRN_GAS7loss\tRN7SL605Ploss\t591\t564\t1.12E-38\n",
    "FAM230Closs\tTXN2_FOXRED2_EIF3D_CACNG2_IFT27_PVALB_NCF4_CSF2RB_TEX33_TST_MPST_KCTD17_RN7SKP214_TMPRSS6_IL2RB_C1QTNF6_SSTR3_RAC2_CYTH4_ELFN2_MFNG_CARD10_CDC42EP1_LGALS2_GGA1_SH3BP1_PDXP_RN7SL385P_LGALS1_NOL12_TRIOBP_H1F0_GCAT_GALR3_ANKRD54_MIR658_MIR659_EIF3L_MICALL1_C22orf23_POLR2F_SOX10_MIR4534_PICK1_SLC16A8_BAIAP2L2_PLA2G6_MAFF_TMEM184B_RN7SL704P_CSNK1E_KCNJ4_KDELR3_DDX17_DMC1_FAM227A_CBY1_TOMM22_JOSD1_GTPBP1_SUN2_DNAL4_NPTXR_CBX6_APOBEC3A_APOBEC3B_APOBEC3C_APOBEC3D_APOBEC3F_APOBEC3G_APOBEC3H_CBX7_PDGFB_RPL3_SNORD83B_SNORD83A_SYNGR1_TAB1_MGAT3_ATF4_RPS19BP1_CACNA1I_ENTHD1_RN7SKP210_GRAP2_FAM83F_TNRC6B_ADSL_SGSM3_MKL1_MCHR1_SLC25A17_MIR4766_ST13_XPNPEP3_DNAJB7_RBX1_EP300_MIR1281_L3MBTL2_CHADL_RANGAP1_ZC3H7B_TEF_TOB2_PHF5A_ACO2_POLR3H_CSDC2_PMM1_DESI1_XRCC6_NHP2L1_C22orf46_MEI1_RNU6ATAC22P_CCDC134_SREBF2_MIR33A_SHISA8_TNFRSF13C_MIR378I_CENPM_LINC00634_WBP2NL_NAGA_FAM109B_SMDT1_NDUFA6_CYP2D6_TCF20_NFAM1_SERHL_RRP7A_SERHL2_RN7SKP80_POLDIP3_RNU12_CYB5R3_ATP5L2_A4GALT_ARFGAP3_PACSIN2_TTLL1_BIK_MCAT_TSPO_TTLL12_SCUBE1_MPPED1_EFCAB6_SULT4A1_PNPLA5_PNPLA3_SAMM50_PARVB_PARVG_KIAA1644_LDOC1L_LINC00207_LINC00229_PRR5_ARHGAP8_PHF21B_NUP50_KIAA0930_MIR1249_UPK3A_FAM118A_SMC1B_RIBC2_FBLN1_ATXN10_MIR4762_WNT7B_LINC00899_MIR3619_MIRLET7A3_MIR4763_MIRLET7B_PPARA_CDPF1_PKDREJ_TTC38_GTSE1_TRMU_CELSR1_GRAMD4_CERKloss\t410\t476\t1.13E-38\n",
    "MYH13_MYH8_MYH4_MYH1_MYH2_MYH3_SCO1_ADPRM_TMEM220_LINC00675_PIRT_RN7SL601P_SHISA6loss\tRN7SL605Ploss\t585\t564\t1.25E-38\n",
    "LINC00670_MYOCD_ARHGAP44_MIR1269B_ELAC2loss\tRN7SL605Ploss\t579\t564\t1.96E-38\n",
    "HS3ST3A1_MIR548H3_COX10_CDRT15_HS3ST3B1_CDRT7_CDRT8_PMP22_MIR4731_TEKT3_RN7SL792P_CDRT4_TVP23C_CDRT1_TRIM16_ZNF286A_TBC1D26_RNA5SP436_ADORA2B_ZSWIM7_TTC19_NCOR1_RN7SL442P_PIGL_MIR1288_CENPV_UBB_TRPV2_ZNF287_ZNF624_CCDC144A_RN7SL620P_USP32P1_KRT16P2_KRT17P1_TBC1D27_TNFRSF13Bloss\tRN7SL605Ploss\t575\t564\t6.06E-38\n",
    "\n",
    "Locations:\n",
    "Gene0Loc\tGene1Loc\n",
    "13:18195297\t22:18077920\n",
    "13:18195297\t22:21754500\n",
    "13:18195297\t22:21919420\n",
    "13:18195297\t22:15528158\n",
    "13:18195297\t22:23573125\n",
    "13:18195297\t22:26169474\n",
    "13:18195297\t22:27978014\n",
    "13:18195297\t22:28056153\n",
    "13:18195297\t22:28642979\n",
    "13:18195297\t22:27748277\n",
    "13:18195297\t22:28883592\n",
    "13:18195297\t22:33436582\n",
    "13:18195297\t22:35066136\n",
    "13:18195297\t22:33162226\n",
    "17:1420689\t11:57528085\n",
    "17:142789\t11:57528085\n",
    "17:1344272\t11:57528085\n",
    "13:18195297\t22:35257452\n",
    "17:5772234\t11:57528085\n",
    "13:18195297\t22:47621043\n",
    "13:18195297\t22:46762617\n",
    "13:18195297\t22:48274364\n",
    "17:6393693\t11:57528085\n",
    "17:7857746\t11:57528085\n",
    "17:11598431\t11:57528085\n",
    "17:12020824\t11:57528085\n",
    "17:9250471\t11:57528085\n",
    "13:18195297\t22:36467036\n",
    "17:10298084\t11:57528085\n",
    "17:12549782\t11:57528085\n",
    "17:13495689\t11:57528085\n",
    "\n",
    "This looks good.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of pairs  94\n",
      "Pair info of triplets initialized\n",
      "36  triplets calculated in  0.0374140739441\n",
      "Including sorting time  0.037712097168\n",
      "19  of type  CooccurringMutuallyExclusiveMutuallyExclusive\n",
      "Triplets written to  /Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-triplet-CooccurringMutuallyExclusiveMutuallyExclusive.txt_Triplets.tsv\n",
      "7  of type  CooccurringCooccurringMutuallyExclusive\n",
      "Triplets written to  /Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-triplet-CooccurringCooccurringMutuallyExclusive.txt_Triplets.tsv\n",
      "10  of type  CooccurringCooccurringCooccurring\n",
      "Triplets written to  /Users/jlu96/conte/jlu/Analyses/CooccurImprovement/LorenzoModel/Binomial/BRCA_wustl-cna-jl-triplet-CooccurringCooccurringCooccurring.txt_Triplets.tsv\n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.433368310598 p_g_m 0.383001049318\n",
      "p_cmm  0.238689552735 expected  227.471143757\n",
      "observed prop 0.495278069255 num overlap  472 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 p_g_m 0.383001049318\n",
      "p_cmm  0.483671709926 expected  460.939139559\n",
      "observed prop 0.495278069255 num overlap  472 out of  953 total \n",
      "Coccur pair :  410 427 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.448058761805 p_g_m 0.383001049318\n",
      "p_cmm  0.239383225667 expected  228.132214061\n",
      "observed prop 0.460650577125 num overlap  439 out of  953 total \n",
      "Coccur pair :  410 427 Mutex:  365\n",
      "p_g_cs 0.397691500525 p_g_m 0.383001049318\n",
      "p_cmm  0.476060025831 expected  453.685204617\n",
      "observed prop 0.460650577125 num overlap  439 out of  953 total \n",
      "Coccur pair :  410 429 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.450157397692 p_g_m 0.383001049318\n",
      "p_cmm  0.2394823218 expected  228.226652676\n",
      "observed prop 0.456453305352 num overlap  435 out of  953 total \n",
      "Coccur pair :  410 429 Mutex:  365\n",
      "p_g_cs 0.396642182581 p_g_m 0.383001049318\n",
      "p_cmm  0.475814487634 expected  453.451206716\n",
      "observed prop 0.456453305352 num overlap  435 out of  953 total \n",
      "Coccur pair :  410 442 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.463798530955 p_g_m 0.383001049318\n",
      "p_cmm  0.240126446666 expected  228.840503673\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 442 Mutex:  365\n",
      "p_g_cs 0.389296956978 p_g_m 0.383001049318\n",
      "p_cmm  0.474095720258 expected  451.813221406\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.400839454355 p_g_m 0.383001049318\n",
      "p_cmm  0.476796640421 expected  454.387198321\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.433368310598 p_g_m 0.383001049318\n",
      "p_cmm  0.238689552735 expected  227.471143757\n",
      "observed prop 0.482686253935 num overlap  460 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.414480587618 p_g_m 0.383001049318\n",
      "p_cmm  0.479988636977 expected  457.429171039\n",
      "observed prop 0.482686253935 num overlap  460 out of  953 total \n",
      "Coccur pair :  410 411 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.431269674711 p_g_m 0.383001049318\n",
      "p_cmm  0.238590456602 expected  227.376705142\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "Coccur pair :  410 411 Mutex:  365\n",
      "p_g_cs 0.420776495278 p_g_m 0.383001049318\n",
      "p_cmm  0.481461866156 expected  458.833158447\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "Coccur pair :  410 476 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.499475341028 p_g_m 0.383001049318\n",
      "p_cmm  0.24181108093 expected  230.445960126\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 476 Mutex:  365\n",
      "p_g_cs 0.401888772298 p_g_m 0.383001049318\n",
      "p_cmm  0.477042178617 expected  454.621196222\n",
      "observed prop 0.450157397692 num overlap  429 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.465897166842 num overlap  444 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.399790136411 p_g_m 0.383001049318\n",
      "p_cmm  0.476551102224 expected  454.15320042\n",
      "observed prop 0.465897166842 num overlap  444 out of  953 total \n",
      "Coccur pair :  410 449 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.471143756558 p_g_m 0.383001049318\n",
      "p_cmm  0.240473283132 expected  229.171038825\n",
      "observed prop 0.451206715635 num overlap  430 out of  953 total \n",
      "Coccur pair :  410 449 Mutex:  365\n",
      "p_g_cs 0.391395592865 p_g_m 0.383001049318\n",
      "p_cmm  0.474586796651 expected  452.281217209\n",
      "observed prop 0.451206715635 num overlap  430 out of  953 total \n",
      "Coccur pair :  410 438 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.459601259182 p_g_m 0.383001049318\n",
      "p_cmm  0.2399282544 expected  228.651626443\n",
      "observed prop 0.447009443861 num overlap  426 out of  953 total \n",
      "Coccur pair :  410 438 Mutex:  365\n",
      "p_g_cs 0.384050367261 p_g_m 0.383001049318\n",
      "p_cmm  0.472868029275 expected  450.643231899\n",
      "observed prop 0.447009443861 num overlap  426 out of  953 total \n",
      "Coccur pair :  410 448 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.470094438615 p_g_m 0.383001049318\n",
      "p_cmm  0.240423735065 expected  229.123819517\n",
      "observed prop 0.461699895068 num overlap  440 out of  953 total \n",
      "Coccur pair :  410 448 Mutex:  365\n",
      "p_g_cs 0.402938090241 p_g_m 0.383001049318\n",
      "p_cmm  0.477287716814 expected  454.855194124\n",
      "observed prop 0.461699895068 num overlap  440 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.464847848898 num overlap  443 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.399790136411 p_g_m 0.383001049318\n",
      "p_cmm  0.476551102224 expected  454.15320042\n",
      "observed prop 0.464847848898 num overlap  443 out of  953 total \n",
      "Coccur pair :  410 433 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.454354669465 p_g_m 0.383001049318\n",
      "p_cmm  0.239680514067 expected  228.415529906\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 433 Mutex:  365\n",
      "p_g_cs 0.399790136411 p_g_m 0.383001049318\n",
      "p_cmm  0.476551102224 expected  454.15320042\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 432 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.453305351522 p_g_m 0.383001049318\n",
      "p_cmm  0.239630966 expected  228.368310598\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 432 Mutex:  365\n",
      "p_g_cs 0.401888772298 p_g_m 0.383001049318\n",
      "p_cmm  0.477042178617 expected  454.621196222\n",
      "observed prop 0.462749213012 num overlap  441 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.439664218258 p_g_m 0.383001049318\n",
      "p_cmm  0.238986841135 expected  227.754459601\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 419 Mutex:  365\n",
      "p_g_cs 0.400839454355 p_g_m 0.383001049318\n",
      "p_cmm  0.476796640421 expected  454.387198321\n",
      "observed prop 0.466946484785 num overlap  445 out of  953 total \n",
      "Coccur pair :  410 424 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.444910807975 p_g_m 0.383001049318\n",
      "p_cmm  0.239234581467 expected  227.990556139\n",
      "observed prop 0.472193074502 num overlap  450 out of  953 total \n",
      "Coccur pair :  410 424 Mutex:  365\n",
      "p_g_cs 0.409233997901 p_g_m 0.383001049318\n",
      "p_cmm  0.478760945994 expected  456.259181532\n",
      "observed prop 0.472193074502 num overlap  450 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.433368310598 p_g_m 0.383001049318\n",
      "p_cmm  0.238689552735 expected  227.471143757\n",
      "observed prop 0.480587618048 num overlap  458 out of  953 total \n",
      "Coccur pair :  410 413 Mutex:  365\n",
      "p_g_cs 0.416579223505 p_g_m 0.383001049318\n",
      "p_cmm  0.48047971337 expected  457.897166842\n",
      "observed prop 0.480587618048 num overlap  458 out of  953 total \n",
      "Coccur pair :  410 405 Mutex:  365\n",
      "p_g_cs 0.430220356768 0.424973767051 p_g_m 0.383001049318\n",
      "p_cmm  0.238293168202 expected  227.093389297\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "Coccur pair :  410 405 Mutex:  365\n",
      "p_g_cs 0.419727177335 p_g_m 0.383001049318\n",
      "p_cmm  0.48121632796 expected  458.599160546\n",
      "observed prop 0.488982161595 num overlap  466 out of  953 total \n",
      "[4.6018466838794629e-66, 2.2732617823899843e-50, 1.5792674494309096e-48, 1.2999042568991835e-45, 3.5065305274658016e-53, 3.5427359777070553e-60, 3.1666592115289563e-63, 6.9401758233186083e-45, 1.2282207405392256e-52, 5.0503605965456466e-46, 1.3083400304917592e-44, 3.1069102022201346e-50, 2.4868579803310521e-52, 4.111061589494067e-51, 3.9558686804913671e-51, 3.5065305274658016e-53, 2.3422524970912992e-55, 2.8066765268020981e-59, 2.5958333653137317e-63]\n",
      "[0.23800684121758911, 0.17354352097606404, 0.12149421845662704, 0.072188557275518123, 0.27972758032205808, 0.43564308425428999, 0.32502595491233122, 0.049110062288141014, 0.25837568090533358, 0.076772550650045832, 0.055986633193093729, 0.17352856009159795, 0.23785989414945108, 0.19964389431079288, 0.19072343568438904, 0.27972758032205808, 0.34864363845427337, 0.5, 0.32499101827171334]\n"
     ]
    }
   ],
   "source": [
    "# Search for triplets\n",
    "import mutex_triangles as met\n",
    "import edgereader as edg\n",
    "reload(chi)\n",
    "reload(met)\n",
    "\n",
    "pairsdict = new_cpairsdict.copy()\n",
    "genedict = edg.get_gene_dict(pairsdict)\n",
    "Triplets, pairsdict_Triplets, sorted_pairs, genesdict_Triplets, sorted_genes = met.getTriplets(pairsdict, genedict, pairsdict, numCases, geneToCases=geneToCases, patientToGenes=patientToGenes)\n",
    "Triplet_dict = met.sort_triplets_by_type(Triplets)\n",
    "\n",
    "for t_type in Triplet_dict:\n",
    "    print len(Triplet_dict[t_type]), \" of type \",  t_type\n",
    "    met.writeTriplets(Triplet_dict[t_type], triplet_file_prefix + t_type + '.txt')\n",
    "\n",
    "# Let's built a nonparametric method to test for significance\n",
    "\n",
    "pvalues = []\n",
    "pvalues_ab = []\n",
    "for i in range(len(Triplets)):\n",
    "    \n",
    "    t_genes = Triplets[i].genes\n",
    "    if Triplets[i].type == 'CooccurringMutuallyExclusiveMutuallyExclusive':\n",
    "        pvalues.append(chi.get_triplet_BinomP(t_genes, geneToCases, patientToGenes,\n",
    "                                     cpairs=Triplets[i].cpairs,\n",
    "                                     mpairs=Triplets[i].mpairs))\n",
    "        pvalues_ab.append(chi.get_triplet_BinomP_ab(t_genes, geneToCases, patientToGenes,\n",
    "                                             cpairs=Triplets[i].cpairs,\n",
    "                                             mpairs=Triplets[i].mpairs))\n",
    "\n",
    "               \n",
    "print pvalues\n",
    "print pvalues_ab\n",
    "\n",
    "\n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Summary of analysis with taking in 80 least mutated patients, using CNAs\n",
    "# Result: \n",
    "1. dominated by co-occurrence among a few patients and between a few genes that weren't filtered out by the MB filter\n",
    "2. not enough gene co-occurrence/mutex within the 80 least mutated\n",
    "\n",
    "\n",
    "Filters:\n",
    "-50 MB\n",
    "-at least 5 x in first cluster, size 80\n",
    "\n",
    "Log:\n",
    "-running with 50MB seems to result in a bunch of pairs between the same several patients (33, 16)\n",
    "\n",
    "TELL LORENZO: no more cooccurrence. Try a bigger partition of the data? Look into mutex?\n",
    "\n",
    "-One significant pair at 0.005. Two genes mutated 3 times, cooccur twice\n",
    "-1276 mutated 3 times, cooccur in 1.\n",
    "\n",
    "-plot p-value distribution\n",
    "-put on figure\n",
    "-run with bigger partitions: size 100\n",
    "\n",
    "Gene0\tGene1\tMutationFrequency0\tMutationFrequency1\tOverlap\t1Freqs0\t1Overlap0\t1CBinomProb0\tBinomProbability\n",
    "TMEM132Dloss\tCCDC91loss\t196\t103\t52\t[3, 3]\t0\t1\t5.62E-09\n",
    "TMEM132Dloss\tBCL2L14loss\t196\t156\t65\t[3, 3]\t0\t1\t1.05E-07\n",
    "TMEM132Dloss\tLRP6loss\t196\t158\t65\t[3, 3]\t0\t1\t2.01E-07\n",
    "TMEM132Dloss\tRERGloss\t196\t159\t65\t[3, 3]\t0\t1\t2.23E-07\n",
    "TMEM132Dloss\tEMP1loss\t196\t170\t68\t[3, 3]\t0\t1\t2.28E-07\n",
    "TMEM132Dloss\tKIAA1467loss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tHEBP1loss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tSNORD88loss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tGSG1loss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tHTR7P1loss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tGPRC5Dloss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tMIR614loss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tGPRC5Aloss\t196\t171\t68\t[3, 3]\t0\t1\t3.76E-07\n",
    "TMEM132Dloss\tETV6loss\t196\t156\t63\t[3, 3]\t0\t1\t4.82E-07\n",
    "TMEM132Dloss\tMANSC1loss\t196\t164\t65\t[3, 3]\t0\t1\t5.88E-07\n",
    "TMEM132Dloss\tDDX47loss\t196\t172\t67\t[3, 3]\t0\t1\t7.00E-07\n",
    "TMEM132Dloss\tAPOLD1loss\t196\t172\t67\t[3, 3]\t0\t1\t7.00E-07\n",
    "TMEM132Dloss\tSTRAPloss\t196\t155\t62\t[3, 3]\t0\t1\t7.66E-07\n",
    "TMEM132Dloss\tPLBD1loss\t196\t165\t65\t[3, 3]\t0\t1\t9.62E-07\n",
    "TMEM132Dloss\tLOH12CR2loss\t196\t165\t65\t[3, 3]\t0\t1\t9.62E-07\n",
    "\n",
    "Significant co-occurrence\n",
    "TMEM132dLOSS, ccdc91loss\n",
    "\n",
    "Can we test more? these were only mutated 3 each in the least mutated guy.\n",
    "\n",
    "\n",
    "Conclusion: 70 too small, not enough mutual exclusivity/co-occurrence\n",
    "Even when rerunning for cluster size 100. not"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Not importing the Fortran bipartite_edge_swap bindings in permute.py.\n"
     ]
    }
   ],
   "source": [
    "\n",
    "__author__ = 'jlu96'\n",
    "import mutex as mex\n",
    "import matplotlib.pyplot as plt\n",
    "import csv\n",
    "import numpy as np\n",
    "from sklearn import mixture\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.cross_validation import KFold\n",
    "from scipy.stats import poisson\n",
    "from scipy import stats\n",
    "import collections\n",
    "import os\n",
    "\n",
    "def partition_EM(patientToGenes, k):\n",
    "    \"\"\"\n",
    "    :param geneToCases:\n",
    "    :param patientToGenes:\n",
    "    :param k: Number of partitions\n",
    "    :return: cohort_list\n",
    "    \"\"\"\n",
    "\n",
    "    # partition the patients, and intersect the geneToCases\n",
    "    return\n",
    "\n",
    "\n",
    "\n",
    "def partition_gene(patientToGenes, genes):\n",
    "    \"\"\"\n",
    "    :param geneToCases:\n",
    "    :param patientToGenes:\n",
    "    :param genes:\n",
    "    :return: cohorts by each gene. Size 2^(#genes)\n",
    "    \"\"\"\n",
    "\n",
    "    cohorts = [patientToGenes.keys()]\n",
    "    for gene in genes:\n",
    "        new_cohorts = []\n",
    "        for cohort in cohorts:\n",
    "            new_cohort_1 = [patient for patient in patientToGenes if gene not in patientToGenes[patient]]\n",
    "            if new_cohort_1:\n",
    "                new_cohorts.append(new_cohort_1)\n",
    "            new_cohort_2 = list(set(cohort).difference(set(new_cohort_1)))\n",
    "            if new_cohort_2:\n",
    "                new_cohorts.append(new_cohort_2)\n",
    "        cohorts = new_cohorts\n",
    "    # print genes\n",
    "    # print cohorts\n",
    "\n",
    "    return cohorts\n",
    "\n",
    "def partition_gene_list(patientToGenes, genes, binary=True):\n",
    "    \"\"\"\n",
    "    :param patientToGenes:\n",
    "    :param genes:\n",
    "    :return: The cohorts, ordered from least to greatest in number of those genes they have.\n",
    "    If binary = True, return just those with, those without.\n",
    "\n",
    "    \"\"\"\n",
    "\n",
    "\n",
    "\n",
    "    gene_set = set(genes)\n",
    "    cohort_dict = {}\n",
    "\n",
    "    for patient in patientToGenes:\n",
    "        num = len(set.intersection(gene_set, patientToGenes[patient]))\n",
    "\n",
    "        # just 0 and 1\n",
    "        if binary:\n",
    "            if num > 0:\n",
    "                num = 1\n",
    "\n",
    "        if num not in cohort_dict:\n",
    "            cohort_dict[num] = []\n",
    "        cohort_dict[num].append(patient)\n",
    "\n",
    "\n",
    "    return cohort_dict\n",
    "\n",
    "\n",
    "def get_patients_gene_mut_num(patients, genes, patientToGenes):\n",
    "    return [set.intersection(patientToGenes[p], genes) for p in patients]\n",
    "\n",
    "def integrate_cohorts(cohort_dict, numCases, num_integrated):\n",
    "    cohorts_int = {}\n",
    "    start_index = 0\n",
    "    num_in_cohort = 0\n",
    "    new_cohort = []\n",
    "    for i in cohort_dict.keys():\n",
    "        num_in_cohort += len(cohort_dict[i])\n",
    "        new_cohort.extend(cohort_dict[i])\n",
    "        if (num_in_cohort > numCases/num_integrated):\n",
    "            cohorts_int[start_index] = new_cohort\n",
    "            start_index = i+1\n",
    "            new_cohort = []\n",
    "            num_in_cohort = 0\n",
    "\n",
    "    if new_cohort:\n",
    "        cohorts_int[start_index] = new_cohort\n",
    "\n",
    "    return cohorts_int\n",
    "\n",
    "def integrate_cohorts_sizes(cohort_dict, sizes):\n",
    "    cohorts_int = {}\n",
    "    size_index = 0\n",
    "    num_in_cohort = 0\n",
    "    new_cohort = []\n",
    "    for i in cohort_dict.keys():\n",
    "        num_in_cohort += len(cohort_dict[i])\n",
    "        new_cohort.extend(cohort_dict[i])\n",
    "        if (num_in_cohort > sizes[size_index]):\n",
    "            cohorts_int[size_index] = new_cohort\n",
    "            size_index += 1\n",
    "            new_cohort = []\n",
    "            num_in_cohort = 0\n",
    "\n",
    "    if new_cohort:\n",
    "        cohorts_int[size_index] = new_cohort\n",
    "\n",
    "    return cohorts_int\n",
    "\n",
    "\n",
    "def draw_partitions_cohorts(geneToCases, patientToGenes, cohort_pairings, title=None, num_bins=50):\n",
    "    # LEFT OF HERE, JLU. Finish this, then above. Make plots in parallel, compare.\n",
    "    # Work with: TP53? Others?\n",
    "\n",
    "    numGenes = len(geneToCases.keys())\n",
    "    numCohorts = len(cohort_pairings)\n",
    "\n",
    "    cohort_frequencies = [[len(patientToGenes[case]) for case in cohort_pair[1]] for cohort_pair in cohort_pairings]\n",
    "    cohort_names = [cohort_pair[0] for cohort_pair in cohort_pairings]\n",
    "\n",
    "    draw_partitions(patientToGenes, cohort_names, cohort_frequencies, title=title, num_bins=num_bins)\n",
    "\n",
    "\n",
    "def draw_partitions(patientToGenes, cohort_names, cohort_frequencies, title=None, num_bins=50):\n",
    "\n",
    "    numCohorts = len(cohort_frequencies)\n",
    "    bins = range(0, max([len(p_gene) for p_gene in patientToGenes.values()]), max([len(p_gene) for p_gene in patientToGenes.values()])/num_bins)\n",
    "\n",
    "    plt.figure()\n",
    "\n",
    "\n",
    "    for i in range(len(cohort_frequencies)):\n",
    "        plt.hist(cohort_frequencies[i], bins, alpha=1.0/numCohorts, label=str(cohort_names[i]))\n",
    "\n",
    "\n",
    "    plt.title(title, fontsize=20)\n",
    "    plt.xlabel('# Somatic Mutations In Tumor', fontsize=20)\n",
    "    plt.ylabel('Number of Samples', fontsize=20)\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "def norm(x, height, center, std):\n",
    "    return(height*np.exp(-(x - center)**2/(2*std**2)))\n",
    "\n",
    "\n",
    "\n",
    "def partition_GMM(patientToGenes, num_components, num_bins, title=None, do_plot=True):\n",
    "    g = mixture.GMM(n_components=num_components)\n",
    "    mut_num_list = [len(patientToGenes[p]) for p in patientToGenes]\n",
    "    obs = np.array([[entry] for entry in mut_num_list])\n",
    "    g.fit(obs)\n",
    "\n",
    "    print \"***********************************\"\n",
    "    print \"COMPONENTS: \", num_components\n",
    "    print \"Weights: \" + str(np.round(g.weights_,2))\n",
    "    print \"Means: \" + str(np.round(g.means_,2))\n",
    "    print \"Covariates: \" + str(np.round(g.covars_,2))\n",
    "\n",
    "    print \"Total log probability: \" + str(sum(g.score(obs)))\n",
    "    print \"AIC: \" + str(g.aic(obs))\n",
    "    print \"BIC: \", g.bic(obs)\n",
    "\n",
    "    score, respon = g.score_samples(obs)\n",
    "\n",
    "    for i in range(num_components):\n",
    "        print \"Model \", np.round(g.means_, 2)[i], \" explains \", np.round(len([in_w for in_w in respon if in_w[i] == max(in_w)])) * 1.0 /len(respon)\n",
    "\n",
    "\n",
    "    # Simulate gaussians\n",
    "    # sim_samples = g.sample(len(patientToGenes))\n",
    "    bins = range(0, max([len(p_gene) for p_gene in patientToGenes.values()]), max([len(p_gene) for p_gene in patientToGenes.values()])/num_bins)\n",
    "    histogram = np.histogram([len(patientToGenes[p]) for p in patientToGenes], bins=bins)\n",
    "\n",
    "    # get the scale of the gaussians from the biggest one\n",
    "    # max_comp = g.weights_.index(max(g.weights_))\n",
    "    # max_mean = g.means_[max_comp]\n",
    "\n",
    "    which_bins = [[bin for bin in bins if bin > mean][0] for mean in g.means_]\n",
    "    print which_bins\n",
    "    print bins\n",
    "    print histogram\n",
    "    print bins.index(which_bins[0]) - 1\n",
    "    bin_heights = [histogram[0][bins.index(which_bin) - 1] for which_bin in which_bins]\n",
    "    # max_height = max(histogram)\n",
    "\n",
    "    if do_plot:\n",
    "        plt.figure()\n",
    "        plt.hist([len(patientToGenes[p]) for p in patientToGenes], bins=bins)\n",
    "        for i in range(num_components):\n",
    "            X = np.arange(0, max(mut_num_list), 1)\n",
    "            Y = norm(X, bin_heights[i], g.means_[i], np.sqrt(g.covars_[i]))\n",
    "            plt.plot(X, Y, label=str(np.round(g.weights_[i], 3)), linewidth=5)\n",
    "        plt.title(\"GMM size \" + str(num_components), fontsize=20)\n",
    "        plt.xlabel('# Somatic Mutations In Tumor', fontsize=20)\n",
    "        plt.ylabel('Number of Samples', fontsize=20)\n",
    "        plt.legend()\n",
    "        plt.show()\n",
    "        # draw_partitions(patientToGenes, ['Original', 'Simulated'], [[len(patientToGenes[p]) for p in patientToGenes], sim_samples],\n",
    "        #                 num_bins=num_bins, title=title)\n",
    "\n",
    "    data = {}\n",
    "    data['Components'] = num_components\n",
    "    data['Weights'] = np.round(g.weights_,2)\n",
    "    data['Means'] = np.round(g.means_,2)\n",
    "    # data['Covariates'] = np.round(g.covars_,2)\n",
    "    # data[\"Total log probability\"] = sum(g.score(obs))\n",
    "    data[\"AIC\"] = g.aic(obs)\n",
    "    data[\"BIC\"] = g.bic(obs)\n",
    "    data['Explained'] = [np.round([len([in_w for in_w in respon if in_w[i] == max(in_w)]) * 1.0 /len(respon) for i in range(num_components)], 2)]\n",
    "\n",
    "    return data\n",
    "\n",
    "def partition_gene_kmeans(geneToCases, patientToGenes, gene_list, num_components, num_bins, title=None, do_plot=True):\n",
    "\n",
    "    # get gene index mapping\n",
    "    giv = getgiv(geneToCases.keys(), gene_list)\n",
    "\n",
    "    # convert patients into vectors\n",
    "    patientToVector = getpatientToVector(patientToGenes, giv)\n",
    "\n",
    "    vectors = patientToVector.values()\n",
    "\n",
    "    print vectors[0]\n",
    "    print \"Length of vectors is \", len(vectors[0])\n",
    "\n",
    "    km = KMeans(num_components)\n",
    "\n",
    "    km.fit(vectors)\n",
    "\n",
    "    clusterToPatient = {}\n",
    "\n",
    "    for patient in patientToVector:\n",
    "        cluster = km.predict(patientToVector[patient])[0]\n",
    "        if cluster not in clusterToPatient:\n",
    "            clusterToPatient[cluster] = set()\n",
    "        clusterToPatient[cluster].add(patient)\n",
    "\n",
    "    # plot patients in each cluster\n",
    "\n",
    "\n",
    "    if do_plot:\n",
    "        bins = range(0, max([len(p_gene) for p_gene in patientToGenes.values()]), max([len(p_gene) for p_gene in patientToGenes.values()])/num_bins)\n",
    "        plt.figure()\n",
    "        for cluster in clusterToPatient:\n",
    "            plt.hist([len(patientToGenes[p]) for p in clusterToPatient[cluster]], bins=bins, label=str(cluster), alpha = 1.0/num_components)\n",
    "        plt.xlabel('# Somatic Mutations In Tumor', fontsize=20)\n",
    "        plt.ylabel('Number of Samples', fontsize=20)\n",
    "        plt.legend()\n",
    "        plt.title(\"Kmeans size \" + str(num_components), fontsize=20)\n",
    "        plt.show()\n",
    "\n",
    "\n",
    "\n",
    "    data = {}\n",
    "    data['Score'] = km.score(vectors)\n",
    "    data['Number'] = num_components\n",
    "    data['% Explained'] = np.round([100 * len(clusterToPatient[cluster]) * 1.0 / len(patientToGenes) for cluster in clusterToPatient], 2)\n",
    "    data['Vector size'] = len(vectors[0])\n",
    "    # data['Covariates'] = np.round(g.covars_,2)\n",
    "    # data[\"Total log probability\"] = sum(g.score(obs))\n",
    "    # data[\"AIC\"] = g.aic(obs)\n",
    "    # data[\"BIC\"] = g.bic(obs)\n",
    "    # data['Explained'] = [np.round([len([in_w for in_w in respon if in_w[i] == max(in_w)]) * 1.0 /len(respon) for i in range(num_components)], 2)]\n",
    "\n",
    "    return data\n",
    "\n",
    "\n",
    "def getgiv(all_genes, gene_list):\n",
    "    \"\"\"\n",
    "    :param all_genes:\n",
    "    :param gene_list:\n",
    "    :return: A list of the genes in common, the gene_index_vector.\n",
    "    \"\"\"\n",
    "    giv = list(set(all_genes).intersection(set(gene_list)))\n",
    "\n",
    "    return giv\n",
    "\n",
    "\n",
    "\n",
    "def getpatientToVector(patientToGenes, gene_index_vector):\n",
    "    patientToVector = {}\n",
    "    for patient in patientToGenes:\n",
    "        patient_genes = patientToGenes[patient]\n",
    "        patientToVector[patient] = []\n",
    "        for gene in gene_index_vector:\n",
    "            patientToVector[patient].append(1 if gene in patient_genes else 0)\n",
    "\n",
    "    return patientToVector\n",
    "\n",
    "\n",
    "def get_cluster_gTC_pTG(geneToCases, patientToGenes, patients):\n",
    "    new_pTG = dict([c for c in patientToGenes.items() if c[0] in patients])\n",
    "    new_genes = set.union(*new_pTG.values())\n",
    "    new_gTC = dict([g for g in geneToCases.items() if g[0] in new_genes])\n",
    "    for g in new_gTC:\n",
    "        new_gTC[g] = new_gTC[g].intersection(patients)\n",
    "    \n",
    "    for g in new_genes:\n",
    "        if g in new_gTC and not new_gTC[g]:\n",
    "            new_gTC.pop(g)\n",
    "    \n",
    "    new_genes = new_genes.intersection(set(new_gTC.keys()))\n",
    "    \n",
    "    return new_genes, new_gTC, new_pTG\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "# 3/12/16-Jlu\n",
    "\n",
    "\n",
    "class PMM:\n",
    "\n",
    "    def __init__(self, filename=None, delimiter='\\t', lam=None, p_k=None, classes=None, patientToGenes=None,\n",
    "                data = None, clusterToPatient = None, do_fit=True):\n",
    "\n",
    "        if filename:\n",
    "            with open(filename, 'rU') as csvfile:\n",
    "                reader = csv.DictReader(csvfile, delimiter=delimiter)\n",
    "                row = reader.next()\n",
    "                print row\n",
    "                self.lam = eval(row['Means'])\n",
    "                self.p_k = eval(row['Probabilities'])\n",
    "                self.classes = eval(row['Classes']) if 'Classes' in row else range(len(self.lam))\n",
    "                self.num_components = len(self.classes)\n",
    "        else:\n",
    "            self.lam = lam\n",
    "            self.p_k = p_k\n",
    "            self.classes = classes\n",
    "            if not classes:\n",
    "                self.classes = range(len(self.lam))\n",
    "            self.num_components = len(self.classes)\n",
    "\n",
    "\n",
    "        self.data = data\n",
    "        self.clusterToPatient = clusterToPatient\n",
    "        print \"Class is \", self.classes, \"Keys are \", self.clusterToPatient.keys()\n",
    "\n",
    "        self.patientToGenes = patientToGenes\n",
    "\n",
    "        if patientToGenes and do_fit:\n",
    "            self.fit_to_data(patientToGenes)\n",
    "\n",
    "    def fit_to_data(self, patientToGenes, min_cluster_size=0):\n",
    "        self.patientToGenes = patientToGenes\n",
    "        self.data, self.clusterToPatient = pmm_fit_to_data(patientToGenes, classes=self.classes, lam=self.lam, p_k=self.p_k,\n",
    "                                                           min_cluster_size=min_cluster_size)\n",
    "        return self.data, self.clusterToPatient\n",
    "\n",
    "\n",
    "    def plot_clusters(self, title):\n",
    "        plot_pmm_clusters(self.patientToGenes, self.clusterToPatient, self.num_components, title=title)\n",
    "\n",
    "\n",
    "    def write_clusters(self, partition_file):\n",
    "        with open(partition_file, 'w') as csvfile:\n",
    "            writer = csv.writer(csvfile)\n",
    "\n",
    "            writer.writerow(['Likelihood', self.data['Likelihood']])\n",
    "            writer.writerow(['BIC', self.data['BIC']])\n",
    "            writer.writerow(['NumComponents', self.data['Number']])\n",
    "            writer.writerow(['Cluster', 'Lambda', 'Probability', 'Patients'])\n",
    "            for k in self.clusterToPatient:\n",
    "                if k != -1:\n",
    "                    lam = self.data['Means'][k]\n",
    "                    p_k = self.data['Probabilities'][k]\n",
    "                else:\n",
    "                    lam = None\n",
    "                    p_k = None\n",
    "                writer.writerow([k, lam, p_k] + list(self.clusterToPatient[k]))\n",
    "\n",
    "    def compare_dna(self, dna_cohort_dict, do_KS=False):\n",
    "\n",
    "        partition_stats_list = []\n",
    "\n",
    "        sizes = [len(self.clusterToPatient[c]) for c in self.clusterToPatient]\n",
    "\n",
    "        # partition by genes\n",
    "        dna_cohorts = integrate_cohorts_sizes(dna_cohort_dict, sizes)\n",
    "\n",
    "        pmm_cluster_list = []\n",
    "        dna_cluster_list = []\n",
    "        \n",
    "        print \"In partition stats Class is \", self.classes, \"Keys are \", self.clusterToPatient.keys()\n",
    "        \n",
    "        for i in range(len(self.classes)):\n",
    "            partition_stats = collections.OrderedDict()\n",
    "            partition_stats['Class'] = self.classes[i]\n",
    "            partition_stats['Mean'] = self.lam[i]\n",
    "            partition_stats['Probability'] = self.p_k[i]\n",
    "\n",
    "\n",
    "            partition_stats['PMM_patients'] = self.clusterToPatient[self.classes[i]]\n",
    "            partition_stats['DNA_patients'] = dna_cohorts[i]\n",
    "\n",
    "            pmm_cluster_list.append(partition_stats['PMM_patients'])\n",
    "            dna_cluster_list.append(partition_stats['DNA_patients'])\n",
    "            \n",
    "            dna_pmn = [len(self.patientToGenes[p]) for p in partition_stats['DNA_patients']]\n",
    "            pmm_pmn = [len(self.patientToGenes[p]) for p in partition_stats['PMM_patients']]\n",
    "\n",
    "            if do_KS:\n",
    "                poisson_cdf.mu = self.lam[i]\n",
    "                partition_stats['KS'] = stats.kstest(dna_pmn, poisson_cdf)\n",
    "\n",
    "            #qq plot of the dna and then the poisson\n",
    "            poisson_q = get_quantiles(dna_pmn, pmm_pmn)\n",
    "            dna_q = get_quantiles(dna_pmn, dna_pmn)\n",
    "\n",
    "            plot_pmm_clusters(self.patientToGenes, {'PMM': partition_stats['PMM_patients'], 'DNA': partition_stats['DNA_patients'] },\n",
    "                              2, num_bins=100, title='DNA VS PMN')\n",
    "\n",
    "            plt.figure()\n",
    "            plt.plot(dna_q, poisson_q, 'bo')\n",
    "            plt.plot([0, 100], [0,100], 'r-', label = 'y=x')\n",
    "            plt.title('QQ for ' + str(self.classes[i]), fontsize=20)\n",
    "            plt.xlabel('DNA_Q', fontsize=20)\n",
    "            plt.ylabel('PMM_Q', fontsize=20)\n",
    "            plt.legend()\n",
    "            plt.show()\n",
    "\n",
    "            partition_stats_list.append(partition_stats)\n",
    "\n",
    "        if do_KS:\n",
    "            self.data['KS_geom_mean'] = mex.prod([partition_stats['KS'][1] for partition_stats in partition_stats_list]) ** (1.0/ len(partition_stats_list))\n",
    "\n",
    "            print \"KS average is \", self.data['KS_geom_mean']\n",
    "            \n",
    "        self.data['CohenKappa'] = cohen_kappa(pmm_cluster_list, dna_cluster_list)\n",
    "\n",
    "\n",
    "        return partition_stats_list\n",
    "\n",
    "\n",
    "\n",
    "def cohen_kappa(cluster_list_1, cluster_list_2):\n",
    "    # assume same categories each\n",
    "    num_agree = 0\n",
    "    prob_agree = 0\n",
    "    total = len(set.union(*[set(c) for c in cluster_list_1]))\n",
    "    \n",
    "    num_classes = len(cluster_list_1)\n",
    "    \n",
    "    cluster_list_1 = [set(c) for c in cluster_list_1]\n",
    "    cluster_list_2 = [set(c) for c in cluster_list_2]\n",
    "    \n",
    "    for k in range(num_classes):\n",
    "        a = cluster_list_1[k]\n",
    "        b = cluster_list_2[k]\n",
    "        num_agree += len(a.intersection(b))\n",
    "        prob_agree += (len(a) * len(b) * 1.0) / (total ** 2)\n",
    "    \n",
    "\n",
    "    obs_agree = num_agree * 1.0 / total\n",
    "    \n",
    "    ck = (obs_agree - prob_agree)/(1.0 - prob_agree)\n",
    "    \n",
    "    print \"Number agreements \", num_agree\n",
    "    print \"Total \", total\n",
    "    print \"Prob agreements \", prob_agree\n",
    "    print \"Cohen kappa \", ck\n",
    "    \n",
    "    return ck\n",
    "        \n",
    "    \n",
    "    \n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def poisson_cdf(x):\n",
    "    if not hasattr(poisson_cdf, 'mu'):\n",
    "        poisson_cdf.mu = 0\n",
    "    print \"X is \", x, \"and mu is \", poisson_cdf.mu\n",
    "    return poisson.cdf(x, poisson_cdf.mu)\n",
    "\n",
    "def get_quantiles(test_dist, base_dist):\n",
    "    return [stats.percentileofscore(base_dist, t) for t in test_dist]\n",
    "\n",
    "def assign_missing(clusterToPatient, patientToGenes):\n",
    "    if -1 not in clusterToPatient:\n",
    "        print \"No missing patients in clusters\"\n",
    "        return clusterToPatient\n",
    "    missing_patients = clusterToPatient[-1]\n",
    "    cluster_means = [(sum([len(patientToGenes[p]) for p in clusterToPatient[c]]) * 1.0 /len(clusterToPatient[c]), c) for c in clusterToPatient if c != -1]\n",
    "    print cluster_means, cluster_means[0][0]\n",
    "    for patient in missing_patients:\n",
    "        num = len(patientToGenes[patient])\n",
    "        correct_cluster = sorted(cluster_means, key=lambda entry: abs(num - entry[0]))[0][1]\n",
    "        clusterToPatient[correct_cluster].add(patient)\n",
    "    clusterToPatient.pop(-1)\n",
    "\n",
    "    return clusterToPatient\n",
    "\n",
    "\n",
    "\n",
    "def best_pmm(patientToGenes, num_components, max_iter=30, rand_num=5, far_rand_num=5, min_cluster_size=0,\n",
    "             plot_clusters=True):\n",
    "\n",
    "    data_record = []\n",
    "    lls_record = []\n",
    "\n",
    "    # Do normal\n",
    "    first_data, lls = partition_pmm(patientToGenes, num_components,  max_iter=max_iter, min_cluster_size=min_cluster_size)\n",
    "\n",
    "    data_record.append(first_data)\n",
    "    lls_record.append(lls)\n",
    "\n",
    "    # Do best rand init\n",
    "    for i in range(rand_num):\n",
    "        data, lls = partition_pmm(patientToGenes, num_components, rand_init=True, max_iter=max_iter, min_cluster_size=min_cluster_size,\n",
    "                                 verbose=False)\n",
    "        data_record.append(data)\n",
    "        lls_record.append(lls)\n",
    "\n",
    "    for i in range(far_rand_num):\n",
    "        data, lls = partition_pmm(patientToGenes, num_components, far_rand_init=True, max_iter=max_iter, min_cluster_size=min_cluster_size,\n",
    "                                 verbose=False)\n",
    "        data_record.append(data)\n",
    "        lls_record.append(lls)\n",
    "\n",
    "    combined_record = zip(data_record, lls_record)\n",
    "\n",
    "    combined_record = sorted(combined_record, key=lambda entry: (-1 * entry[0]['Missing'], entry[0]['Likelihood']), reverse=True)\n",
    "\n",
    "    data_record, lls_record = zip(*combined_record)\n",
    "\n",
    "    best_data = data_record[0]\n",
    "\n",
    "    if (best_data['Likelihood'] > first_data['Likelihood'] + 10):\n",
    "        print \"First data not best!\"\n",
    "        best_data['IsFirst'] = False\n",
    "    else:\n",
    "        best_data['IsFirst'] = True\n",
    "\n",
    "\n",
    "    clusterToPatient = pmm_to_cluster(patientToGenes, best_data['Classes'], best_data['Means'], best_data['Probabilities'])\n",
    "\n",
    "    if plot_clusters:\n",
    "        plot_pmm_clusters(patientToGenes, clusterToPatient, num_components)\n",
    "\n",
    "    plot_likelihoods(lls_record)\n",
    "\n",
    "    return best_data, clusterToPatient\n",
    "    # Return clusters\n",
    "\n",
    "\n",
    "def pmm_to_cluster(patientToGenes, classes, lam, p_k):\n",
    "    clusterToPatient = {}\n",
    "\n",
    "    for k in classes:\n",
    "        clusterToPatient[k] = set()\n",
    "\n",
    "    clusterToPatient[-1] = set()\n",
    "\n",
    "\n",
    "    for patient in patientToGenes:\n",
    "        d = len(patientToGenes[patient])\n",
    "\n",
    "        max_class = -1\n",
    "        max_ll = -np.inf\n",
    "        for k in classes:\n",
    "            if (np.log(p_k[k]) + np.log(poisson(lam[k]).pmf(d))) > -np.inf:\n",
    "                if (np.log(p_k[k]) + np.log(poisson(lam[k]).pmf(d))) > max_ll:\n",
    "                    max_class = k\n",
    "                    max_ll = (np.log(poisson(lam[k]).pmf(d)))\n",
    "\n",
    "\n",
    "        clusterToPatient[max_class].add(patient)\n",
    "\n",
    "    missing_clusters = set()\n",
    "    for cluster in clusterToPatient:\n",
    "        if not clusterToPatient[cluster]:\n",
    "            print '**********NO PATIENTS IN CLUSTER ', lam[cluster], p_k[cluster]\n",
    "            missing_clusters.add(cluster)\n",
    "            #clusterToPatient[cluster].add('NO PATIENTS IN CLUSTER')\n",
    "    for cluster in missing_clusters:\n",
    "        clusterToPatient.pop(cluster)\n",
    "            \n",
    "    return clusterToPatient\n",
    "\n",
    "\n",
    "\n",
    "def pmm_cross_validate(num_components, patientToGenes, num_folds, kf_random_state=None, max_iter=30, rand_num=5, far_rand_num=5, min_cluster_size=0):\n",
    "    \"\"\"\n",
    "    :return: The average likelihood of the model when applied to a new test set, and its BIC\n",
    "    \"\"\"\n",
    "\n",
    "    kf = KFold(len(patientToGenes), n_folds=num_folds, random_state=kf_random_state)\n",
    "\n",
    "    lls = []\n",
    "    missing_patients = []\n",
    "    bics = []\n",
    "    for train_index, test_index in kf:\n",
    "\n",
    "        train_patientToGenes = dict([patientToGenes.items()[x] for x in train_index])\n",
    "        test_patientToGenes = dict([patientToGenes.items()[x] for x in test_index])\n",
    "        best_data, _ = best_pmm(train_patientToGenes, num_components, max_iter=max_iter, rand_num=rand_num,\n",
    "                                               far_rand_num=far_rand_num, min_cluster_size=min_cluster_size)\n",
    "\n",
    "        test_stats, test_cluster = pmm_fit_to_data(test_patientToGenes, best_data['Classes'], best_data['Means'], best_data['Probabilities'])\n",
    "\n",
    "        plot_pmm_clusters(test_patientToGenes, test_cluster, num_components, title='Test clusters size ' + str(num_components))\n",
    "\n",
    "        lls.append(test_stats['Likelihood'])\n",
    "        missing_patients.append(test_stats['Missing'])\n",
    "        bics.append(test_stats['BIC'])\n",
    "\n",
    "    return sum(lls) * 1.0/len(lls), sum(missing_patients) * 1.0 / len(missing_patients), sum(bics) * 1.0/ len(bics)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def pmm_fit_to_data(patientToGenes, classes, lam, p_k, data=None, min_cluster_size=0):\n",
    "    \"\"\"\n",
    "    :param patientToGenes:\n",
    "    :param lam:\n",
    "    :param p_k:\n",
    "    :param data:\n",
    "    :return: data, clusterToPatient\n",
    "    \"\"\"\n",
    "\n",
    "    if not data:\n",
    "        data = collections.OrderedDict()\n",
    "\n",
    "\n",
    "    D = [len(patientToGenes[p]) for p in patientToGenes]\n",
    "    numCases = len(D)\n",
    "    num_components = len(lam)\n",
    "\n",
    "    ll_kd = np.array([ [np.log(p_k[k]) + np.log(poisson(lam[k]).pmf(d)) for d in D] for k in classes])\n",
    "    likelihood_sums = np.zeros(numCases)\n",
    "\n",
    "    for i in range(numCases):\n",
    "        likelihood_sums[i] = sum([(np.exp(ll_kd[k][i]) if ll_kd[k][i] > -np.inf else 0) for k in range(num_components)] )\n",
    "\n",
    "    # complete log likelihood\n",
    "\n",
    "    ll = sum(np.log(np.array([ls for ls in likelihood_sums if ls > 0])))\n",
    "\n",
    "    clusterToPatient = pmm_to_cluster(patientToGenes, classes, lam, p_k)\n",
    "\n",
    "    print \"LL:\", np.round(ll), \"Missing patients: \", len(clusterToPatient[-1]) if -1 in clusterToPatient else 0\n",
    "\n",
    "    data['Number'] = num_components\n",
    "    data['OriginalNumber'] = num_components\n",
    "    mp = zip(*sorted(zip(list(np.round(lam, 1)), list(np.round(p_k, 2))), key = lambda entry: entry[0]))\n",
    "\n",
    "    data['Means'], data['Probabilities'] =  list(mp[0]), list(mp[1])   \n",
    "    data['Likelihood'] = np.round(ll)\n",
    "    data['Classes'] = classes\n",
    "    data['AIC'] = np.round(2 * (len(p_k) + len(lam)) - 2 * ll)\n",
    "    data['BIC'] = np.round(-2 * ll + (len(p_k) + len(lam)) * np.log(numCases))\n",
    "    data['Missing'] = len(clusterToPatient[-1]) if -1 in clusterToPatient else 0\n",
    "    data['MinClusterSize'] = min([len(clusterToPatient[c]) if c != -1 else np.inf  for c in clusterToPatient])\n",
    "    data['MoreThanMin'] = 1 if data['MinClusterSize'] > min_cluster_size else 0\n",
    "    data['Merged'] = False\n",
    "    data['MergeHistory'] = set()\n",
    "\n",
    "    return data, clusterToPatient\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def partition_pmm(patientToGenes, num_components, diff_thresh=0.01, num_bins=50, max_iter=100, by_iter=True,\n",
    "                  rand_init=False, far_rand_init=False, do_plot=False, get_best=True, min_cluster_size=0,\n",
    "                 verbose=True):\n",
    "\n",
    "\n",
    "    # get the whole data distribution\n",
    "\n",
    "\n",
    "    # D = [1,2,3,4,5, 100, 150, 200, 1000]\n",
    "    D = [len(patientToGenes[p]) for p in patientToGenes]\n",
    "    numCases = len(D)\n",
    "    data = collections.OrderedDict()\n",
    "\n",
    "    # print \"D is \", D\n",
    "\n",
    "    # get the lambdas at equal-spaced intervals\n",
    "\n",
    "\n",
    "    lam = [np.percentile(D, (i + 1) * 100.0 / (num_components + 1)) for i in range(num_components)]\n",
    "    p_k = [1.0 / num_components for i in range(num_components)]\n",
    "    classes = range(num_components)\n",
    "\n",
    "    if rand_init:\n",
    "        old_lam = lam\n",
    "        old_p_k = p_k\n",
    "        #random sample  in a range centered at the quartiles\n",
    "        lam = [np.random.uniform(l - 0.5 * old_lam[0], l + 0.5 * old_lam[0]) for l in old_lam]\n",
    "        rand_freq = [2**np.random.uniform(-1, 1) * pk for pk in old_p_k]\n",
    "        p_k = list(np.array(rand_freq)/sum(rand_freq))\n",
    "        classes = range(num_components)\n",
    "\n",
    "    if far_rand_init:\n",
    "        lam = [np.random.uniform(min(D), max(D)) for l in lam]\n",
    "        rand_freq = [np.random.uniform(0, 1) for l in lam]\n",
    "        p_k = list(np.array(rand_freq)/sum(rand_freq))\n",
    "\n",
    "    if verbose:\n",
    "        print \"Initial Lambda is \", lam\n",
    "        print \"Initial p_k is\", p_k\n",
    "\n",
    "    data['Initial Means'] = np.round(lam,1)\n",
    "    data['Initial p_k'] = np.round(p_k, 2)\n",
    "\n",
    "    ll = -3e100\n",
    "    num_iter = 0\n",
    "\n",
    "    # stupid inital values\n",
    "    p_k_d= np.zeros(num_components)\n",
    "    lam_prev = np.zeros(num_components)\n",
    "    p_k_prev = np.zeros(num_components)\n",
    "\n",
    "    # for the best values\n",
    "    ll_best = -np.inf\n",
    "    p_k_best = None\n",
    "    lam_best = None\n",
    "    missing_best = numCases\n",
    "\n",
    "    lls = []\n",
    "\n",
    "    while 1:\n",
    "\n",
    "\n",
    "        # We have the log-likelihood of data d and class k in matrix\n",
    "        #            data 1 data 2 data 3\n",
    "        # clsss 1   ll_11   ll_12\n",
    "        # class 2\n",
    "        ll_kd = np.array([ [np.log(p_k[k]) + np.log(poisson(lam[k]).pmf(d)) for d in D] for k in classes])\n",
    "\n",
    "        \n",
    "\n",
    "        # Likelihood_sums: the total likelihood of each data, summed across class k\n",
    "        likelihood_sums = np.zeros(numCases)\n",
    "\n",
    "        for i in range(numCases):\n",
    "            likelihood_sums[i] = sum([(np.exp(ll_kd[k][i]) if ll_kd[k][i] > -np.inf else 0) for k in range(num_components)] )\n",
    "\n",
    "            \n",
    "        missing_new = len([x for x in likelihood_sums if x == 0])\n",
    "        # complete log likelihood\n",
    "\n",
    "        ll_new = sum(np.log(np.array([ls for ls in likelihood_sums if ls > 0])))\n",
    "\n",
    "        if num_iter == 0:\n",
    "            data['Initial LL'] = np.round(ll_new)\n",
    "\n",
    "        if verbose:\n",
    "            print \"ll_new is \", ll_new, \"missing is \", missing_new\n",
    "\n",
    "\n",
    "        if ll_new > ll_best or missing_new < missing_best:\n",
    "            ll_best = ll_new\n",
    "            p_k_best = p_k\n",
    "            lam_best = lam\n",
    "            missing_best = missing_new\n",
    "\n",
    "        # When we break out of the loop, take previous value since it might have jumped out\n",
    "        if (by_iter):\n",
    "            if num_iter > max_iter:\n",
    "                break\n",
    "            elif abs(ll_new - ll) < diff_thresh:\n",
    "                break\n",
    "        else:\n",
    "            if abs(ll_new - ll) < diff_thresh:\n",
    "\n",
    "                p_k_d = p_k_d_prev\n",
    "                lam = lam_prev\n",
    "                p_k = p_k_prev\n",
    "\n",
    "            break\n",
    "\n",
    "        p_k_d_prev = p_k_d\n",
    "        lam_prev = lam\n",
    "        p_k_prev = p_k\n",
    "\n",
    "\n",
    "        # Calculate p_k_d. This is p(data d | class k) * p(class k)/sum(p(data|class i) *p(class i);\n",
    "        # i.e. prob of this class given this data\n",
    "\n",
    "        p_k_d = np.zeros(ll_kd.shape)\n",
    "\n",
    "        for i in range(numCases):\n",
    "            # Use max class likelihood to divide all the likelihoods by\n",
    "            max_val = np.amax(ll_kd, axis=0)[i]\n",
    "\n",
    "            # sum the likekhoods for every class, make this the denominator of probability\n",
    "            denom = sum([(np.exp(ll_kd[k][i] - max_val) if ll_kd[k][i] > -np.inf else 0) for k in range(num_components)])\n",
    "\n",
    "            for k in range(num_components):\n",
    "                p_k_d[k][i] = (np.exp(ll_kd[k][i] - max_val) / denom if ll_kd[k][i] > -np.inf else 0)\n",
    "                # print \"numerator is \", np.exp(ll_kd[k][i] - max), \" prob is \", p_k_d[k][i]\n",
    "\n",
    "        # print \"p_k_d is \", p_k_d\n",
    "\n",
    "        # sum probabilities of each data being each class over all data\n",
    "        Z_k = p_k_d.sum(axis=1)\n",
    "\n",
    "\n",
    "        # see derivation\n",
    "\n",
    "        lam = [sum([p_k_d[k][i] * D[i] for i in range(numCases)]) * 1.0 / Z_k[k] for k in classes]\n",
    "        p_k = Z_k * 1.0 / numCases\n",
    "\n",
    "        p_k = p_k/p_k.sum()\n",
    "\n",
    "\n",
    "        # print \"New lambda is \", lam\n",
    "        # print \"New p_k is \", p_k\n",
    "\n",
    "\n",
    "        ll = ll_new\n",
    "\n",
    "        lls.append(ll)\n",
    "        num_iter += 1\n",
    "\n",
    "\n",
    "\n",
    "    if get_best:\n",
    "        p_k = p_k_best\n",
    "        lam = lam_best\n",
    "        ll = ll_best\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "    data, clusterToPatient = pmm_fit_to_data(patientToGenes, classes, lam, p_k, data=data, min_cluster_size=min_cluster_size)\n",
    "    # plot patients in each cluster\n",
    "\n",
    "    if do_plot:\n",
    "        plot_pmm_clusters(patientToGenes, clusterToPatient, num_components, num_bins=100)\n",
    "\n",
    "\n",
    "    # clusterToPatient = pmm_to_cluster(patientToGenes, classes, lam, p_k)\n",
    "\n",
    "    #\n",
    "    #\n",
    "    #\n",
    "    #\n",
    "    # data['Number'] = num_components\n",
    "    # data['Means'] = np.round(lam, 1)\n",
    "    # data['Probabilities'] = np.round(p_k, 2)\n",
    "    # data['Likelihood'] = np.round(ll)\n",
    "    # data['Classes'] = classes\n",
    "    # data['AIC'] = np.round(2 * (len(p_k) + len(lam)) - 2 * ll)\n",
    "    # data['BIC'] = np.round(-2 * ll + (len(p_k) + len(lam)) * np.log(numCases))\n",
    "    # data['Missing'] = len(clusterToPatient[-1]) if -1 in clusterToPatient else 0\n",
    "    # data['MinClusterSize'] = min([len(clusterToPatient[c]) if c != -1 else np.inf  for c in clusterToPatient])\n",
    "    # data['MoreThanMin'] = 1 if data['MinClusterSize'] > min_cluster_size else 0\n",
    "\n",
    "    return data, lls\n",
    "\n",
    "\n",
    "\n",
    "def sort_data_by_means(data):\n",
    "    \"\"\" Sort in ascending order. Don't need to change cluster labels\"\"\"\n",
    "    data_items = data.items()\n",
    "    mean_indices = ((i, data['Means'][i]) for i in range(len(data['Means'])))\n",
    "    mean_indices = sorted(mean_indices, key=lambda entry: min(entry[1]) if isinstance(entry[1], list)\n",
    "                         else entry[1])\n",
    "    \n",
    "    conversion_array = [m[0] for m in mean_indices] # this should map to the correct index now. these are new clusters\n",
    "    \n",
    "    new_data = collections.OrderedDict()\n",
    "    \n",
    "    for key in data:\n",
    "        value = data[key]\n",
    "        if isinstance(value, np.ndarray):\n",
    "            new_value = np.zeros(len(value))\n",
    "            for i in range(len(conversion_array)):\n",
    "                new_value[i] = value[conversion_array[i]]\n",
    "            new_data[key] = new_value\n",
    "        if isinstance(value, list):\n",
    "            new_value = [value[conversion_array[i]] for i in range(len(conversion_array))]\n",
    "            new_data[key] = new_value\n",
    "            \n",
    "        else:\n",
    "            new_data[key] = value\n",
    "    \n",
    "    return new_data\n",
    "    \n",
    "\n",
    "def merge_clusters(data, clusterToPatient, patientToGenes,\n",
    "                  missing_limit=0.5, min_cluster_size=30):\n",
    "    \"\"\"Merge adjacent clusters. Choosse to merge those clusters that\n",
    "    are the most similar, as measured by the likelihood of one within\n",
    "    another.\n",
    "    missing_limit is the limit on number of patients that can't\n",
    "    be explained by one cluster. Clusters will be sorted first\n",
    "    by those who are below the minimum cluster size,\n",
    "    less missing patients in their merging\n",
    "    cluster, then by those that have the highest likelihood\n",
    "    \"\"\"\n",
    "    # get the likelihood of each cluster rel. to other ones\n",
    "    # only look at adjacent clusters! sort them\n",
    "    \n",
    "    data = sort_data_by_means(data)\n",
    "    \n",
    "    print \"****************************************\"\n",
    "    print \"Begin merging.\"\n",
    "    # first go forward\n",
    "\n",
    "    \n",
    "    classes = data['Classes']\n",
    "    p_k = data['Probabilities']\n",
    "    lam = data['Means']\n",
    "    \n",
    "    \n",
    "    all_list = []\n",
    "    \n",
    "    for i in range(len(lam) - 1):\n",
    "        from_index, to_index = i, i + 1\n",
    "        from_class, to_class = classes[from_index], classes[to_index]\n",
    "        patients = clusterToPatient[from_class]\n",
    "        p = [len(patientToGenes[patient]) for patient in patients]\n",
    "        \n",
    "        #check if we're dealing with merged clusters. if so... add the likelihoods of the individual\n",
    "        # underlying poissons?\n",
    "        if isinstance(p_k[from_index], list):\n",
    "            clust_probs = p_k[from_index]\n",
    "            clust_means = lam[from_index]\n",
    "            clust_size = len(clust_means)\n",
    "            \n",
    "            from_ll = [max([np.log(clust_probs[x]) + \n",
    "                           np.log(poisson(clust_means[x]).pmf(d)) for x in range(clust_size)])\n",
    "                          for d in p]\n",
    "        else:\n",
    "            from_ll = [np.log(p_k[from_index]) + np.log(poisson(lam[from_index]).pmf(d)) for d in p]\n",
    "            \n",
    "        if isinstance(p_k[to_index], list):\n",
    "            clust_probs = p_k[to_index]\n",
    "            clust_means = lam[to_index]\n",
    "            clust_size = len(clust_means)\n",
    "            \n",
    "            to_ll = [max([np.log(clust_probs[x]) + \n",
    "                           np.log(poisson(clust_means[x]).pmf(d)) for x in range(clust_size)])\n",
    "                          for d in p]\n",
    "        else:\n",
    "            to_ll = [np.log(p_k[to_index]) + np.log(poisson(lam[to_index]).pmf(d)) for d in p]\n",
    "            \n",
    "        missing = np.isinf(from_ll) ^ np.isinf(to_ll)\n",
    "        \n",
    "        missing_indices = np.where(missing)[0]\n",
    "        good_indices = np.where(~missing)[0]\n",
    "        \n",
    "        missing_num = len(missing_indices)\n",
    "        \n",
    "        ll_diffs = [to_ll[j] - from_ll[j] for j in good_indices]\n",
    "        \n",
    "        ll_diffs_total = sum(ll_diffs)\n",
    "        \n",
    "        all_list.append([(from_index, to_index), missing_num, ll_diffs_total, missing_num > missing_limit * len(p),\n",
    "                        len(patients) < min_cluster_size])\n",
    "        \n",
    "    # now go backwards\n",
    "    for i in reversed(range(1, len(lam))):\n",
    "        from_index, to_index = i, i - 1\n",
    "        from_class, to_class = classes[from_index], classes[to_index]\n",
    "        patients = clusterToPatient[from_class]\n",
    "        p = [len(patientToGenes[patient]) for patient in patients]\n",
    "        \n",
    "                #check if we're dealing with merged clusters. if so... add the likelihoods of the individual\n",
    "        # underlying poissons?\n",
    "        if isinstance(p_k[from_index], list):\n",
    "            clust_probs = p_k[from_index]\n",
    "            clust_means = lam[from_index]\n",
    "            clust_size = len(clust_means)\n",
    "            \n",
    "            from_ll = [max([np.log(clust_probs[x]) + \n",
    "                           np.log(poisson(clust_means[x]).pmf(d)) for x in range(clust_size)])\n",
    "                          for d in p]\n",
    "        else:\n",
    "            from_ll = [np.log(p_k[from_index]) + np.log(poisson(lam[from_index]).pmf(d)) for d in p]\n",
    "            \n",
    "        if isinstance(p_k[to_index], list):\n",
    "            clust_probs = p_k[to_index]\n",
    "            clust_means = lam[to_index]\n",
    "            clust_size = len(clust_means)\n",
    "            \n",
    "            to_ll = [max([np.log(clust_probs[x]) + \n",
    "                           np.log(poisson(clust_means[x]).pmf(d)) for x in range(clust_size)])\n",
    "                          for d in p]\n",
    "        else:\n",
    "            to_ll = [np.log(p_k[to_index]) + np.log(poisson(lam[to_index]).pmf(d)) for d in p]\n",
    "        \n",
    "        \n",
    "        missing = np.isinf(from_ll) ^ np.isinf(to_ll)\n",
    "        \n",
    "        missing_indices = np.where(missing)[0]\n",
    "        good_indices = np.where(~missing)[0]\n",
    "        \n",
    "        missing_num = len(missing_indices)\n",
    "        \n",
    "        ll_diffs = [to_ll[j] - from_ll[j] for j in good_indices]\n",
    "        \n",
    "        ll_diffs_total = sum(ll_diffs)\n",
    "        \n",
    "        \n",
    "        all_list.append([(from_index, to_index), missing_num, ll_diffs_total, missing_num < missing_limit * len(p),\n",
    "                        len(patients) < min_cluster_size])\n",
    "        \n",
    "    \n",
    "    # sort by the cluster that's below the min size, then byminimum missing, then by maximum likelihood ratio\n",
    "    all_list = sorted(all_list, key=lambda entry: (entry[4], entry[3], entry[2]), reverse=True)\n",
    "    \n",
    "    print \"Possible merged clusters is \", all_list\n",
    "    print \"Best cluster is \", all_list[0]\n",
    "    \n",
    "\n",
    "    (from_index, to_index), missing_num, ll_diffs_total, more_than_missing, cluster_too_small = all_list[0]\n",
    "\n",
    "    # calculate the new AIC, BIC, make new cluster to patient, make new classes..new means? update probabilities\n",
    "    \n",
    "    # Record merge history\n",
    "    new_data = data\n",
    "    if 'MergeHistory' not in new_data:\n",
    "        new_data['MergeHistory'] = set()\n",
    "    \n",
    "    new_data['MergeHistory'].add((str([lam[from_index], lam[to_index]]),\n",
    "                  str([p_k[from_index], p_k[to_index]]),\n",
    "                  (len(clusterToPatient[classes[from_index]]), len(clusterToPatient[classes[to_index]])),\n",
    "                  missing_num, ll_diffs_total, ('Num classes befpre', len(classes), ('Cluster too small?', cluster_too_small))))\n",
    "        \n",
    "    new_clusterToPatient = clusterToPatient\n",
    "    moved_patients = new_clusterToPatient[classes[from_index]]\n",
    "    new_clusterToPatient[classes[to_index]] = new_clusterToPatient[classes[to_index]].union(moved_patients)\n",
    "    new_clusterToPatient.pop(classes[from_index])\n",
    "\n",
    "    \n",
    "    print \"MERGING the probs and likelihoods\"\n",
    "    if not isinstance(p_k[from_index], list):\n",
    "        p_k[from_index] = [p_k[from_index]]\n",
    "        lam[from_index] = [lam[from_index]]\n",
    "    if not isinstance(p_k[to_index], list):\n",
    "        p_k[to_index] = [p_k[to_index]]\n",
    "        lam[to_index] = [lam[to_index]] \n",
    "    p_k[to_index].extend(p_k[from_index])\n",
    "    lam[to_index].extend(lam[from_index])\n",
    "    new_data['Probabilities'] = p_k\n",
    "    new_data['Means'] = lam\n",
    "    \n",
    "    \n",
    "    print \"MERGING: HERE ARE OLD VALUES\", new_data\n",
    "    #remove all the old values\n",
    "    new_data['Merged'] = True\n",
    "    new_data['Number'] -= 1\n",
    "    for key in new_data:\n",
    "        value = new_data[key]\n",
    "        if isinstance(value, np.ndarray):\n",
    "            value = list(value)\n",
    "            value = value[0: from_index] + value[from_index + 1 :]\n",
    "            value = np.array(value)\n",
    "            new_data[key] = value\n",
    "        elif isinstance(value, list):\n",
    "            value = value[0: from_index] + value[from_index + 1 :]\n",
    "            new_data[key] = value\n",
    "\n",
    "    print \"New classe:\", new_data['Classes'], \"VS NEW KEYS\", new_clusterToPatient.keys()\n",
    "            \n",
    "    # integrate the old patients to the new ones\n",
    "\n",
    "    \n",
    "    \n",
    "    new_data['MinClusterSize'] = min(len(new_clusterToPatient[c]) for c in new_clusterToPatient)\n",
    "    \n",
    "    print \"MERGING: HERE ARE NEW VALUES\", new_data\n",
    "    \n",
    "    plot_pmm_clusters(patientToGenes, clusterToPatient, new_data['Number'], title='Merging')\n",
    "    \n",
    "    print \"End merging.\"\n",
    "    print \"****************************************\"    \n",
    "    \n",
    "    return new_data, new_clusterToPatient\n",
    " \n",
    "    \n",
    "#     data['Number'] = num_components\n",
    "#     data['Means'], data['Probabilities'] = zip(*sorted(zip(list(np.round(lam, 1)), list(np.round(p_k, 2))), key = lambda entry: entry[0]))\n",
    "#     data['Likelihood'] = np.round(ll)\n",
    "#     data['Classes'] = classes\n",
    "#     data['AIC'] = np.round(2 * (len(p_k) + len(lam)) - 2 * ll)\n",
    "#     data['BIC'] = np.round(-2 * ll + (len(p_k) + len(lam)) * np.log(numCases))\n",
    "#     data['Missing'] = len(clusterToPatient[-1]) if -1 in clusterToPatient else 0\n",
    "#     data['MinClusterSize'] = min([len(clusterToPatient[c]) if c != -1 else np.inf  for c in clusterToPatient])\n",
    "#     data['MoreThanMin'] = 1 if data['MinClusterSize'] > min_cluster_size else 0\n",
    "\n",
    "def backward_selection(data, clusterToPatient, patientToGenes, min_cluster_size = 30,\n",
    "                       max_components = 10):\n",
    "    \"\"\"Merge clusters until a criterion is satisfied. Missing patients are assumed to\n",
    "    be assigned already.\n",
    "    \"\"\"\n",
    "    \n",
    "\n",
    "    merged_data = data\n",
    "    merged_cluster = clusterToPatient\n",
    "    \n",
    "    while (merged_data['Number'] > max_components or merged_data['MinClusterSize'] < min_cluster_size):\n",
    "        merged_data, merged_cluster = merge_clusters(merged_data, merged_cluster, patientToGenes,\n",
    "                                                    min_cluster_size = min_cluster_size)\n",
    "    \n",
    "    return merged_data, merged_cluster\n",
    "    \n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def plot_pmm_clusters(patientToGenes, clusterToPatient, num_components, num_bins=100, title=None):\n",
    "    D = [len(patientToGenes[p]) for p in patientToGenes]\n",
    "\n",
    "    bins = range(0, max(list(D)), max(list(D))/num_bins)\n",
    "    plt.figure()\n",
    "    for cluster in clusterToPatient:\n",
    "        plt.hist([len(patientToGenes[p]) for p in clusterToPatient[cluster]], bins=bins, label=str(cluster), alpha = 1.0/num_components)\n",
    "    plt.xlabel('# Somatic Mutations In Tumor', fontsize=20)\n",
    "    plt.ylabel('Number of Samples', fontsize=20)\n",
    "    plt.legend()\n",
    "    if not title:\n",
    "        plt.title(\"Cluster size \" + str(num_components), fontsize=20)\n",
    "    else:\n",
    "        plt.title(title, fontsize=20)\n",
    "    plt.show()\n",
    "\n",
    "def plot_likelihoods(ll_record):\n",
    "    plt.figure()\n",
    "    for i in range(len(ll_record)):\n",
    "        plt.plot(ll_record[i], label=str(i))\n",
    "    plt.title(\"Log-likelihood change in EM\", fontsize=20)\n",
    "    plt.legend(loc=4)\n",
    "    plt.show()\n",
    "\n",
    "# If there are any patients that aren't assigned, i.e. in cluster -1\n",
    "# Throw them out?\n",
    "def load_patient_cohorts(partitionfile, patientToGenes, add_to_closest=True, delimiter='\\t'):\n",
    "    clusterToProp = {}\n",
    "\n",
    "    with open(partitionfile, 'rU') as csvfile:\n",
    "        reader = csv.reader(csvfile, delimiter=delimiter)\n",
    "        for row in reader:\n",
    "            if (row[0] == 'Cluster'): break\n",
    "        # reader = csv.DictReader(csvfile, delimiter=delimiter)\n",
    "        # print \"begun dict reader\\n\"\n",
    "        for row in reader:\n",
    "            c = eval(row[0])\n",
    "            clusterToProp[c] = {}\n",
    "            clusterToProp[c]['Mean'] = eval(row[1]) if row[1] else 0\n",
    "            clusterToProp[c]['Probability'] = eval(row[2]) if row[2] else 0\n",
    "            clusterToProp[c]['Patients'] = set(row[3:]) if row[3] else set()\n",
    "\n",
    "\n",
    "    if -1 in clusterToProp:\n",
    "        if add_to_closest:\n",
    "            other_cs = clusterToProp.keys()\n",
    "            other_cs.remove(-1)\n",
    "            print \"Removed \", clusterToProp[-1]\n",
    "            for patient in clusterToProp[-1]:\n",
    "                sims = [(abs(len(patientToGenes[patient]) - clusterToProp[c]['Mean']), c) for c in other_cs]\n",
    "                sims = sorted(sims, key = lambda entry: entry[0])\n",
    "                best_c = sims[0][1]\n",
    "                clusterToProp[best_c]['Patients'].add(patient)\n",
    "            print \"completed\"\n",
    "\n",
    "        clusterToProp.pop(-1)\n",
    "\n",
    "    sorted_clusters = sorted(clusterToProp.keys(), key = lambda entry: clusterToProp[entry]['Mean'])\n",
    "    \n",
    "    oldclusterToProp = clusterToProp.copy()\n",
    "    clusterToProp = {}\n",
    "    cohort_dict = {}\n",
    "    \n",
    "    for i in range(len(sorted_clusters)):\n",
    "        cohort_dict[i] = oldclusterToProp[sorted_clusters[i]]['Patients']\n",
    "        clusterToProp[i] = oldclusterToProp[sorted_clusters[i]]\n",
    "    \n",
    "    min_cohort = cohort_dict[0]\n",
    "    \n",
    "    \n",
    "    \n",
    "    return cohort_dict, clusterToProp, min_cohort\n",
    "\n",
    "# INDEX BY LOSSES\n",
    "%matplotlib inline\n",
    "def run_partitions(mutationmatrix = None, #'/Users/jlu96/maf/new/OV_broad/OV_broad-cna-jl.m2',\n",
    "        patientFile = None, #'/Users/jlu96/maf/new/OV_broad/shared_patients.plst',\n",
    "        out_file = None, #'/Users/jlu96/conte/jlu/Analyses/CancerMutationDistributions/OV_broad-cna-jl-PMM-crossval.txt',\n",
    "        partition_file = None, #'/Users/jlu96/maf/new/OV_broad/OV_broad-cna-jl.ppf',\n",
    "        load_pmm_file = None, #'/Users/jlu96/conte/jlu/Analyses/CancerMutationDistributions/OV_broad-cna-jl-PMM.txt',\n",
    "        dna_pmm_comparison_file = None, #'/Users/jlu96/conte/jlu/Analyses/CancerMutationDistributions/OV_broad-cna-jl-PMM-dnacomp.txt',\n",
    "        cluster_matrix = None, # '/Users/jlu96/maf/new/OV_broad/OV_broad-cna-jl-cluster.m2',\n",
    "        min_cluster_size = 15,\n",
    "        num_init = 9,\n",
    "        minComp = 2,\n",
    "        maxComp = 5,\n",
    "        do_plot = True,\n",
    "        do_gmm = False,\n",
    "        do_dna = False,\n",
    "        num_integrated = 4,\n",
    "        do_kmeans = False,\n",
    "        do_pmm = True,\n",
    "        do_cross_val = False,\n",
    "        do_pmm_dna = True,\n",
    "        do_back_selection = True,\n",
    "        write_cluster_matrices = True,\n",
    "        rand_num = 3,\n",
    "        far_rand_num = 3,\n",
    "        kf_random_state = 1,\n",
    "        kf_num_folds = 5,\n",
    "\n",
    "        geneFile = None,\n",
    "        minFreq = 0,\n",
    "        dna_gene_file = '/Users/jlu96/conte/jlu/Analyses/CancerGeneAnalysis/DNADamageRepair_loss.txt',\n",
    "       out_dir = '/Users/jlu96/conte/jlu/Analyses/CancerMutationDistributions/',\n",
    "        write_all_partitions = True):\n",
    "    \n",
    "    mutationmatrix_list = mutationmatrix.split('/')\n",
    "    matrix_dir = '/'.join(mutationmatrix_list[:-1]) + '/'\n",
    "    prefix = (mutationmatrix_list[-1]).split('.m2')[0]\n",
    "    \n",
    "\n",
    "    if not patientFile:\n",
    "        patientFile = matrix_dir + 'shared_patients.plst'\n",
    "        \n",
    "    if not out_file:\n",
    "        if do_cross_val:\n",
    "            out_file = out_dir + prefix + '-PMM-crossval-kf' + str(kf_num_folds) + '.txt'\n",
    "        else:\n",
    "            out_file = out_dir + prefix + '-PMM-comparisons.txt'\n",
    "    \n",
    "    if not partition_file:\n",
    "        partition_file = matrix_dir + prefix + '.ppf'\n",
    "        \n",
    "    \n",
    "    if not load_pmm_file:\n",
    "        load_pmm_file = out_dir + prefix + '-PMM.txt'\n",
    "    \n",
    "    if not dna_pmm_comparison_file:\n",
    "        dna_pmm_comparison_file = out_dir + prefix + '-PMM-dnacomp.txt'\n",
    "        \n",
    "    if not cluster_matrix:\n",
    "        cluster_matrix = matrix_dir + prefix + '-cluster.m2'\n",
    "\n",
    "    \n",
    "    numGenes, numCases, genes, patients, geneToCases, patientToGenes = mex.load_mutation_data(mutationmatrix, patientFile, geneFile, minFreq)\n",
    "\n",
    "    p_gene_list = []\n",
    "\n",
    "    with open(dna_gene_file, 'rU') as row_file:\n",
    "        reader = csv.reader(row_file, delimiter='\\t')\n",
    "        for row in reader:\n",
    "            p_gene_list.append(row[0])\n",
    "        dna_cohort_dict = partition_gene_list(patientToGenes, p_gene_list, binary=not bool(num_integrated))\n",
    "\n",
    "\n",
    "    if do_kmeans:\n",
    "        datas = []\n",
    "        for i in np.arange(minComp, maxComp, 1):\n",
    "            datas.append(partition_gene_kmeans(geneToCases, patientToGenes, p_gene_list, i, num_bins=50, title=None, do_plot=True))\n",
    "\n",
    "        with open(out_file, 'w') as csvfile:\n",
    "            writer = csv.DictWriter(csvfile, fieldnames=datas[0].keys())\n",
    "            writer.writeheader()\n",
    "            for row in datas:\n",
    "                writer.writerow(row)\n",
    "\n",
    "\n",
    "    if do_dna:\n",
    "        cohort_dict = partition_gene_list(patientToGenes, p_gene_list, binary=not bool(num_integrated))\n",
    "        # Make new cohorts over this\n",
    "        if num_integrated:\n",
    "            cohort_dict = integrate_cohorts(cohort_dict, numCases, num_integrated)\n",
    "\n",
    "\n",
    "        cohort_pairings = [(key, cohort_dict[key]) for key in cohort_dict]\n",
    "        draw_partitions_cohorts(geneToCases, patientToGenes, cohort_pairings, title='DNADamageGenes',\n",
    "                        num_bins=100 if mutationmatrix[-9:] == 'cna-jl.m2' else 50)\n",
    "\n",
    "\n",
    "    if do_gmm:\n",
    "        datas = []\n",
    "        for i in np.arange(minComp, maxComp, 1):\n",
    "            datas.append(partition_GMM(patientToGenes, i, num_bins=50, title='GMM size ' + str(i), do_plot=do_plot))\n",
    "\n",
    "        with open(out_file, 'w') as csvfile:\n",
    "            writer = csv.DictWriter(csvfile, fieldnames=datas[0].keys())\n",
    "            writer.writeheader()\n",
    "            for row in datas:\n",
    "                writer.writerow(row)\n",
    "\n",
    "\n",
    "    if do_pmm:\n",
    "        datas = []\n",
    "        clusters = []\n",
    "\n",
    "        partition_stats_list = []\n",
    "        for num_components in np.arange(minComp, maxComp, 1):\n",
    "            best_data, clusterToPatient = best_pmm(patientToGenes, num_components, rand_num=rand_num, far_rand_num=far_rand_num,\n",
    "                                                   min_cluster_size=min_cluster_size)\n",
    "\n",
    "            if do_back_selection:\n",
    "                # assign the missing data\n",
    "                clusterToPatient = assign_missing(clusterToPatient, patientToGenes)\n",
    "                best_data, clusterToPatient = backward_selection(best_data, clusterToPatient, patientToGenes, min_cluster_size = min_cluster_size,\n",
    "                       max_components = maxComp)\n",
    "            \n",
    "            if do_pmm_dna:\n",
    "                print \"cfirst lasses are \", best_data['Classes'], \"clusterToPatient is \", clusterToPatient.keys()\n",
    "                pmm = PMM(lam=best_data['Means'], p_k=best_data['Probabilities'], patientToGenes=patientToGenes,\n",
    "                         data=best_data, clusterToPatient=clusterToPatient, classes=best_data['Classes'],\n",
    "                          do_fit=False)\n",
    "\n",
    "                partition_stats_list.extend(pmm.compare_dna(dna_cohort_dict))\n",
    "\n",
    "                best_data = pmm.data\n",
    "\n",
    "\n",
    "            if do_cross_val:\n",
    "            #cross validate each of the components\n",
    "                print \"*******************************************************************************************************\"\n",
    "                print \"BEGINNING CROSS VALIDATION for \", num_components\n",
    "                print \"*******************************************************************************************************\"\n",
    "                best_data['TestLL'], best_data['TestMissing'], best_data['TestBIC'] = pmm_cross_validate(num_components, patientToGenes,\n",
    "                                                                                                         num_folds=kf_num_folds,\n",
    "                                                                                                     kf_random_state=kf_random_state,\n",
    "                                                                                   rand_num=rand_num, far_rand_num=far_rand_num,\n",
    "                                                                                   min_cluster_size=min_cluster_size)\n",
    "                best_data['TestFolds'] = kf_num_folds\n",
    "\n",
    "                print \"*******************************************************************************************************\"\n",
    "                print \"EMDING CROSS VALIDATION  for \", num_components\n",
    "                print \"*******************************************************************************************************\"\n",
    "\n",
    "            datas.append(best_data)\n",
    "            clusters.append(clusterToPatient)\n",
    "            \n",
    "            if write_all_partitions:\n",
    "                with open(partition_file + str(num_components), 'w') as csvfile:\n",
    "                    writer = csv.writer(csvfile, delimiter='\\t')\n",
    "\n",
    "                    writer.writerow(['Likelihood', best_data['Likelihood']])\n",
    "                    writer.writerow(['BIC', best_data['BIC']])\n",
    "                    writer.writerow(['NumComponents', best_data['Number']])\n",
    "                    writer.writerow(['Cluster', 'Mean', 'Probability', 'Patients'])\n",
    "                    if 'Merged' in best_data and best_data['Merged']:\n",
    "                        for k in range(len(clusterToPatient)):\n",
    "                            lam = best_data['Means'][k]\n",
    "                            p_k = best_data['Probabilities'][k]\n",
    "                            writer.writerow([best_data['Classes'][k] , lam, p_k]  + list(clusterToPatient[best_data['Classes'][k]]))\n",
    "                        \n",
    "                    else:\n",
    "                        for k in clusterToPatient:\n",
    "                            if k != -1:\n",
    "                                lam = best_data['Means'][k]\n",
    "                                p_k = best_data['Probabilities'][k]\n",
    "                            else:\n",
    "                                lam = None\n",
    "                                p_k = None\n",
    "                            writer.writerow([k, lam, p_k] + list(clusterToPatient[k]))\n",
    "\n",
    "        # get the best BIC\n",
    "        combined = zip(datas, clusters)\n",
    "        if do_cross_val:\n",
    "            combined = sorted(combined, key=lambda entry: ( -1 * entry[0]['MoreThanMin'], np.round(entry[0]['TestMissing']), -1 * entry[0]['TestLL'], entry[0]['TestBIC'], entry[0]['BIC']))\n",
    "        else:\n",
    "            combined = sorted(combined, key=lambda entry: ( -1 * entry[0]['MoreThanMin'], entry[0]['BIC']))\n",
    "\n",
    "        datas, clusters = zip(*combined)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "        with open(out_file, 'w') as csvfile:\n",
    "            writer = csv.DictWriter(csvfile, fieldnames=datas[-1].keys(), delimiter='\\t', extrasaction='ignore')\n",
    "            print datas\n",
    "            writer.writeheader()\n",
    "            for row in datas:\n",
    "                writer.writerow(row)\n",
    "\n",
    "\n",
    "        best_data = datas[0]\n",
    "        clusterToPatient = clusters[0]\n",
    "\n",
    "        # code to parition by best clusters\n",
    "        with open(partition_file, 'w') as csvfile:\n",
    "            writer = csv.writer(csvfile, delimiter='\\t')\n",
    "\n",
    "            writer.writerow(['Likelihood', best_data['Likelihood']])\n",
    "            writer.writerow(['BIC', best_data['BIC']])\n",
    "            writer.writerow(['NumComponents', best_data['Number']])\n",
    "            writer.writerow(['Cluster', 'Mean', 'Probability', 'Patients'])\n",
    "            if 'Merged' in best_data and best_data['Merged']:\n",
    "                for k in range(len(clusterToPatient)):\n",
    "                    lam = best_data['Means'][k]\n",
    "                    p_k = best_data['Probabilities'][k]\n",
    "                    writer.writerow([best_data['Classes'][k] , lam, p_k]  + list(clusterToPatient[best_data['Classes'][k]]))\n",
    "                        \n",
    "            else:\n",
    "                for k in clusterToPatient:\n",
    "                    if k != -1:\n",
    "                        lam = best_data['Means'][k]\n",
    "                        p_k = best_data['Probabilities'][k]\n",
    "                    else:\n",
    "                        lam = None\n",
    "                        p_k = None\n",
    "                    writer.writerow([k, lam, p_k] + list(clusterToPatient[k]))\n",
    "\n",
    "        if write_cluster_matrices:\n",
    "            for cluster in clusterToPatient:\n",
    "                with open(cluster_matrix + str(cluster), 'w') as csvfile:\n",
    "                    writer = csv.writer(csvfile, delimiter='\\t')\n",
    "                    for patient in clusterToPatient[cluster]:\n",
    "                        writer.writerow('\\t'.join([patient] + list(patientToGenes[patient])))\n",
    "\n",
    "\n",
    "        if do_pmm_dna:\n",
    "            with open(dna_pmm_comparison_file, 'w') as csvfile:\n",
    "                writer = csv.DictWriter(csvfile, fieldnames=partition_stats_list[0].keys(), delimiter='\\t')\n",
    "                writer.writeheader()\n",
    "                print \"header written\"\n",
    "                for row in partition_stats_list:\n",
    "                    writer.writerow(row)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}