Uncovering rare variants in HLA genes: evolutionary analysis and impact on genetic load

Abstract

Genome-wide studies with hundreds and even thousands of individuals and millions of markers are becoming increasingly common and are profoundly changing our understanding of human genetic diversity. Important themes which are currently under intense research include a detailed understanding of the demographic history of our species (How intensely did populations expand? How long ago were continents colonized?) as well as the selective forces acting on the human genome (Which genes are selected? How common are deleterious variants?). Whereas genome-wide studies provide broad explanations regarding human genetic variation and evolutionary forces, they leave out details concerning selective regimes acting on specific genes. The analysis of individual genes can provide information concerning important aspects of the evolutionary process, including the timing of selection, its intensity, the relationship between selective and demographic processes, and the effect of deleterious variants (i.e., the intensity of load) in explaining phenotypic variation. Here we propose to study the Human Leukocyte Antigen (HLA) genes in a unprecedented scale using two large genome-wide data sets comprising 7,604 individuals from publicly available next-generation sequencing (NGS) databases. The HLA genes code for proteins of central importance in the immune system, involved in the response to pathogenic infections. Their variability is associated with the efficiency of the immune system to respond to infections. From an evolutionary perspective, HLA genes are a classical example of balancing selection, with host-pathogen interactions driving the diversity at these loci, through selective regimes that may be caused by heterozygote advantage, frequency-dependent selection, selection that varies over space and/or time, or a mixture of these. The high variability at HLA loci results in important challenges to the accuracy and applicability of bioinformatic tools used in genome-wide analyses, including read mapping and variant calling. For this reason it is paramount to develop adequate tools which can accurately quantify variation within these highly polymorphic genes. In this project we propose to develop analytical tools which will provide reliable HLA polymorphism data based on large samples generated by NGS. With this data we will address questions concerning how genetic drift and natural selection have interacted to shape extant HLA variation. Specifically, we will address the following questions. (1) What is the variability of common and rare segregating sites of HLA genes and how is it distributed among populations in a worldwide scale? We will investigate the hypothesis that human populations harbor a very large reservoir of currently undetected rare variants within HLA genes, which constitute a resource for future adaptive evolution. (2) How does balancing selection on HLA loci affect variation and genetic load at linked sites? We will use our large dataset to explore the effects of linkage in driving putatively deleterious variants to higher than expected frequencies, and evaluate how this phenomenum contributes to the overall degree of load in the MHC region. (3) How are genetic and phenotypic (e.g. expression levels) variation related? We will test the hypothesis that expression levels determine the exposure of variants to selection. This will be done by comparing functional variation in high- and low-expressing HLA haplotypes.Our investigation will require extending hlaTX, a bioinformatic tool that quantifies HLA expression and provides genotype calls based on RNAseq data, to settings where exome data is used and where sample sizes are very large. We will integrate our empirical analyses with the use of computer simulations, which will provide a framework to test hypotheses of evolutionary processes that can account for features of the data we will quantify, including levels of polymorphism and genetic load. (AU)