An organism's phenotype depends on a multitude of genetic variants, spread over a linear genome. This is widely understood, and yet in practice, has hardly been incorporated into population genetic analysis. Recent developments in theory, computation, and sequencing technology now make it possible to obtain and analyse whole haploid genomes on a large scale. This proposal is to develop and apply a theoretical analysis of genetic variation that is spread over continuous linear genomes. Theory and methods will be developed in close interaction with empirical data from artificial selection experiments and from an intensely studied hybrid zone in Antirrhinum; for both, we have a known pedigree, and phased whole-genome sequence. Population structure will be analysed by following blocks of genome through pedigrees, and across two-dimensional landscapes. Selection on discrete loci will be analysed by finding its effect on surrounding haplotypes, by analysing how favoured alleles become disentangled from heterogeneous backgrounds, and by seeing how haplotype blocks flow past selected clines. The contribution of variants that are spread across the genome to GWA, to selection response, and to hybrid zones will be modelled, and the overall effect of inherited fitness variance on haplotype structure will be determined. This work will establish a new framework for population genomics that goes beyond the current focus on individual loci. It will help bridge the distinct communities within genomics, quantitative genetics, and population genetics, which currently tackle these problems largely in isolation. The project will develop better tools for inferring selection and population structure from DNA sequence data, and more fundamentally, will give us a deeper understanding of how the abundant variation that is carried on linear genomes is shaped by evolution.
