Epigenomewide association

This is furnished with ewas.sh and ewas.subs, along with a few other files as follows,

Files	Description
ewas.sh	EWAS imputation
ewas.subs	subroutine called by ewas.sh
get_weight.qsub	SGE script for EWAS weight generation
get_weight_subs	subroutine callable from get_weight.qsub and parallel
CpG.lst	list of probe IDs with weights
weights/	directory containing weights for all probes as specfied in CpG.lst
EWAS/	directory containing PLINK binary files for each probe
EWAS.pheno	PLINK phenotype file with header for all probes
EWAS.bim	PLINK .bim file sorted by SNP IDs

This implementation used the same idea as TWAS. Data from 1000Genomes imputation were scaled down to those in HapMap II to make the weight generation more tenable to sample size. Note that weights were obtained for all probes so it is possible to impute for only subset(s) of them. The file EWAS.bim was generated in order to make it easier to align strands for SNPs as in GWAS with those in the reference panel.

Additional information is now available from EWAS-fusion.