Test for two-SNP interaction adjusting for long range linkage disequilibrium

Thursday, October 11, 2012 — Poster Session III
10:00 a.m. – Noon	Natcher Conference Center, Building 45	NHGRI	GEN/GENOM-13

Authors

• Qing Li
• Silke Szymczak
• Joan Bailey-Wilson

Abstract

For case-parent trios, we propose a new test for two-SNP interaction that adjusts for linkage disequilibrium (LD). For family-based data, most existing tests for interaction depend on the assumption that the markers are in linkage equilibrium. However, in dense marker panels, we often observe two or more markers are correlated although they are even further apart physically. Without adjustment for long range LD, tests for interaction can have inflated Type-I error rates. Our test utilizes the conditional logistic regression framework for the genotypic TDT test. Instead of using observed case genotypes and untransmitted genotypes generated for pseudo controls to fit the model, we use the probabilities that two alleles at a pair of loci were transmitted together to the case and the probability they were not simultaneously transmitted to weight the relative risk of the interaction between two alleles. To adjust for LD between the loci, we can choose to either rely on haplotype reconstruction or the conditional probability of one allele given the other based on data from founders. We will present simulation results evaluating the Type-I error and power of the test compared to alternative tests, including PTDT, Cordell’s test and case-only test implemented in PLINK.

back to top