Skip to content

Latest commit

 

History

History
60 lines (41 loc) · 3.38 KB

COOKBOOK.md

File metadata and controls

60 lines (41 loc) · 3.38 KB

StrainFLAIR cookbook

Prerequisite

Datasets for this cookbook are available in the folder data/manuscript_simulated_datasets/. It contains 8 complete genomes (fasta files) of Escherichia coli and several mixtures of simulated reads with 0.1% of errors as described in the manuscript (fastq.gz files).

For this cookbook, only the dataset mix_o104_iai39_k12_100k_e0.001.fastq.gz is used. It corresponds to a mix of 50% of E. coli O104:H4 2011C-3493, 33% of IAI39 and 17% of K-12 MG1655. All commands presented can be used for the other datasets, by simply changing the name of the fastq file used as shown in the third step below.

Full commands: The full commands are proposed at the end of the section.

Computation time is approximately 3 minutes for the indexation step, 30 minutes for the mapping step and 1 minute for the query step.

Consider 7 reference genomes:

First, create a file of file (fof) used as input for StrainFLAIR. Each line of the input fof contains exactly one fasta name. The folder contains 8 genomes, however the strain BL21DE3 is only used as an unknown strain be to queried and is not indexed. We may create this fof as follows:

find data/manuscript_simulated_dataset/references/* -not -name "*bl21*" > data/list_fasta_sim.txt

Second, run StrainFLAIR indexation step, using the input fof:

./StrainFLAIR.sh index -i data/list_fasta_sim.txt -o myproject

The reference graph and its associated additionnal files have been created in the new directory myproject.

Third, query the variation graph by mapping any reads on it. Reads can be compressed in a fastq.gz format.

./StrainFLAIR.sh query -g myproject/graphs/all_graphs -f1 data/manuscript_simulated_dataset/mix_o104_iai39_k12_100k_e0.001.fastq.gz -t 24 -p myproject/graphs/dict_clusters.pickle -d myproject -o manuscriptsim

That's it! Mapping files are available in mapping/, and abundance tables are available in results/.

The final result is a csv table containing each reference genome in line identified by their accession number. Columns contained the proportion of detected genes and the estimated strain-level abundance according to different computation methods. Here we found close estimations compared to what is expected: 44.7% for O104:H4 2011C-3493 (NC_018658.1), 35% for IAI39 (NC_011750.1), 20.3% for K-12 MG1655, and 0 for all the absent strains.

Here is the complete output:

,detected_genes,mean_abund,mean_abund_nz,median_abund,median_abund_nz
NC_000913.3,0.9787234042553191,20.29226977437424,21.05505540046613,19.25596456585335,19.939248305563503
NC_002695.2,0.2019113460756405,0.0,0.0,0.0,0.0
NC_018658.1,0.9935683046050939,44.67232236815524,44.1192172117696,46.3219304819112,45.76611823931458
NC_011750.1,0.9885301614273577,35.03540785747052,34.825727387764275,34.42210495223545,34.29463345512191
NZ_CP028116.1,0.22883087400681043,0.0,0.0,0.0,0.0
NZ_CP007592.1,0.20748792270531402,0.0,0.0,0.0,0.0
NC_013654.1,0.15242438790206433,0.0,0.0,0.0,0.0

Full commands:

find data/manuscript_simulated_dataset/references/* -not -name "*bl21*" > data/list_fasta_sim.txt
./StrainFLAIR.sh index -i data/list_fasta_sim.txt -o myproject
./StrainFLAIR.sh query -g myproject/graphs/all_graphs -f1 data/manuscript_simulated_dataset/mix_o104_iai39_k12_100k_e0.001.fastq.gz -t 24 -p myproject/graphs/dict_clusters.pickle -d myproject -o manuscriptsim