Illustrating, Quantifying, and Correcting for Bias in Post-Hoc Analysis of Gene-Based Rare Variant Tests of Association

• Activity Number: 561
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): Kelsey Grinde*
• Companies: University of Washington
• Keywords: rare variants ; winners curse ; gene-based tests

Abstract

To date, gene-based rare variant testing approaches have focused on aggregating information across sets of variants to maximize statistical power in identifying genes that show significant association with a disease. Follow-up, post-hoc analysis to identify causal variant(s) in those genes and estimate their effect is crucial for planning replication studies and characterizing the genetic architecture of the locus. However, we illustrate that straightforward, post-hoc single-marker association statistics suffer from bias introduced by conditioning on gene-based test significance, a phenomenon akin to "winner's curse," which has been well documented in genetic association studies with common variants. We will illustrate the ramifications of this bias on power and type I error of single-marker association tests, outline parameters of genetic architecture that affect this bias, and propose a bootstrap resampling method to correct for this bias. We demonstrate that our bootstrap-corrected estimates offer improved performance (i.e., bias, power, type I error) in many situations. To conclude, we offer suggestions regarding the future use of post-hoc rare variant analysis methods.