Genome-Wide Quantification of Expression QTLs for Mendelian Randomization

• Activity Number: 149
• Type: Invited
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Nancy Zhang*+ and Dylan S Small and Yang Jiang
• Companies: University of Pennsylvania and University of Pennsylvania and University of Pennsylvania

Abstract

High throughput technologies for profiling DNA variation is making possible their use as instruments to estimate the causal effects of expressed gene and protein markers on observed clinical phenotypes. For this analysis technique, called "Mendelian Randomization", to be feasible on the genome-wide scale, there needs to be an existence of good, strong instruments for gene transcripts at the genome level. Existing studies on cis- gene expression mapping suggests that cis-linked SNP markers are abundant and strong. We re-examine these results in a statistically rigorous way, quantify the genome-wide strength of cis-linkage, and evaluate their suitability as instruments for causal studies. This is joint work with Yang Jiang and Dylan Small.