Selection strategies for sib-pair association studies

Abstract

Quantitative-trait association studies are a common approach in behavior genetics research. When a large number of sibships have already been phenotyped but one can only afford to genotype a subset of the available sample, it is desirable to optimize the choice of who to genotype. This poster presents a novel sib-pair sampling strategy for genotyping in QTL association studies, both when only phenotype information is available and also when using a previously genotyped sample from a linkage study with a positive finding to be confirmed and fine-mapped. We show how the genetic effect of a diallelic additive trait locus can be modeled within the maximum-likelihood variance components framework proposed by Fulker et al. (1999) and sib pairs rank-ordered by their informativeness for association. The performance of our method was investigated via simulation. We demonstrate increased power when selecting sib pairs for genotyping based on identity by descent information from a previous linkage study, as compared to using only the phenotypic information. An R-script implementing the selection approaches (including traditional phenotype-based ones) is available for download at https://statgen. hku.hk/jshkwan.