CMStatistics 2016: Start Registration
View Submission - CMStatistics
B0580
Title: Adaptive surrogate confounding adjustment in genomic mediation analysis and an application to GTEx data Authors:  Lin Chen - University of Chicago (United States) [presenting]
Abstract: Recent studies of the effects of genetic variation on expression of distal genes (trans-eQTLs) have revealed that many trans-eQTL effects are mediated by a local (cis-) gene transcript near the eQTL. When conducting mediation analyses in the genome, millions of mediation trios consisting of SNP, cis-gene and trans-gene transcripts are considered and each trio may suffer from different confounding effects. Instead of adjusting for the same set of confounders or surrogate variables in all mediation tests, we propose an adaptive surrogate confounder adjustment algorithm that adaptively selects surrogate confounders for different mediation trios. We applied the proposed methods to a multitissue expression data from National Institute of Health Common Funds Genotype-Tissue Expression (GTEx) program, which has collected transcriptome data in a wide variety of tissue types and genetic data from post-mortem donors. We examined the patterns of cis-gene mediating trans-effect within each tissue type and across different tissue types, with the proposed adaptive surrogate confounding adjustment.