Presentation: Accounting for Gender in Statistical Tests for Hardy-Weinberg Equilibrium
Speaker: Jan Graffelman, PhD, Associate Professor, Department of Statistics and Operations Research, Universitat Politècnica de Catalunya, Barcelona, Spain
Abstract: The Hardy-Weinberg law, as stated more than a century ago, is a cornerstone principle of modern genetics, and refers to the stability of allele and genotype frequencies in a biological population over the generations. Hardy-Weinberg proportions are often assumed in, among others, models in genetic epidemiology, relatedness estimation and in calculations done in forensic genetics. In association studies, genetic variants are typically tested for equilibrium by exact procedures on a genome-wide scale, mainly for quality control purposes with the aim of identifying variants susceptible to genotyping errors. The Hardy-Weinberg law has been a topic of considerable research interest, and there are hundreds of research papers dedicated to the law. Research related to the principle continues as new kinds of genetic data arise.
The X-chromosome is special in many types of genetic analysis for the simple fact that males only have one copy. For decades, Hardy-Weinberg equilibrium on the X-chromosome has therefore been tested by using females only. In recent work we question this approach, and propose novel asymptotic and exact procedures tailored for X-chromosomal variants, that consider potential differences in allele frequencies between the sexes. Insights gleaned from the X-chromosomal work can be taken back to the analysis of the autosomes, where accounting for gender also leads to novel methodological developments that provide new ways to look at genetic data.
In this talk I will given an overview of current statistical techniques used for both autosomal and X-chromosomal inference on Hardy-Weinberg proportions, illustrating them with data from the 1000 genomes project. Considerable part of the talk is joint work with Bruce Weir.