Improved detection of evolutionary selection highlights potential bias from different sequencing strategies in complex genomic-regions

Abstract

Balancing selection occurs when different evolutionary pressures impact the fitness of multiple alleles, resulting in increased allelic diversity in the population. A new statistical method was developed to test for selection, improving inference by using efficient Bayesian techniques to test for density and strength of linkage disequilibrium. Evolutionary simulation studies showed that the method consistently outperformed existing methods. Using this methodology, we tested for novel signals of balancing selection genome wide in 500 samples from phased trios. Several novel signals of selection appeared in CYP2A7, GPC6, and CNR2 across multiple ancestries. Additionally, tests in SIRPA demonstrate dramatically strong selection signal, significantly higher than previously observed. Well-known signals around olfactory genes and the MHC, containing HLA genes associated with the immune response, also demonstrated strong signatures of selection. So, utilizing data from the 17th IHIW, a follow up analysis was then performed by leveraging over seven thousand HLA typed samples by NGS; in contrast, the genome wide scan did not include a detailed characterization of the HLA genes. The strongest signals observed in the IHIW samples were in DQA1 and DQB1 in or around exon 2, the portion of the gene responsible for antigen presentation and most likely to be under environmental and evolutionary pressure. Our new statistical approach and analysis suggest novel evolutionary pressure in new regions and additionally highlight the importance of improved sequencing and characterization of variation across the extended MHC and other critical regions.