Results of Sarkar and Park et al. (2017) submitted.
All the tissue and SNP effect sizes can be bound in
result/stat/chr1/50/combined.txt.gz
(...)
result/stat/chr22/50/combined.txt.gz
Note that individual by tissue gene expression
matrix Y was regressed on cis-regulatory genotype matrix X with
factored effect size matrix.
Y ~ X * sum_k (Theta[, k] * t(Theta[, k])) + confounders + errors
Each row contains
ENSEMBL.ID: unique gene IDChromosome: chromosome name (1to22)TSS: transcription start site (provided by GTEx v6p)Tissue.idx: comma-separate tissue indexesTissue.names: comma-separate tissue namesTissue.theta: comma-separate tissue effect sizesTissue.se: comma-separate tissue effect size standard errorsTissue.lodds: comma-separate tissue PIP log-oddsSNP.names: comma-separate SNP namesSNP.theta: comma-separate SNP effect sizesSNP.se: comma-separate SNP effect size standard errorsSNP.lodds: comma-separate SNP PIP log-oddsk: factor indexpip: posterior inclusion probability cutoff