Fig 1.
A host-pathogens arms race can lead to significant modifications to the human genome of a host population over time. Human proteins that interact directly with pathogen proteins are often the target of strong positive selection. Random variation and novel mutations can naturally lead to increased fitness in certain individuals in a population with an endemic pathogen. Genetic resistance that has arisen due to positive selection acting on protective variants may be detected with a GWAS.
Fig 2.
Worldwide levels of HIV-1 selection.
A p-value for selection in human genes that interact with HIV-1 was produced for each population and is represented by a colored circle on the map. Significance is portrayed by a white to red gradient with lighter colors representing higher p-values and brighter red colors representing lower p-values. Only SNPs (filtered for LD) within genes that interact with Y. Pestis were included in determining the p-values. The most significant p-values were found in East Asia and Africa.
Fig 3.
Overlap of selection targets associated with HIV-1.
Each population was associated with a positive selection score for each gene interacting with Yersinia pestis. If gene scores were correlated for any population pair, there was significant overlap in the genes undergoing selection across both populations. We performed pairwise correlations among the 15 most selected populations to uncover the extent of overlap in selection targets. Kendall’s’ Tau-b coefficient ranged from -0.12 to 0.15. We identified 4 population pairs with significantly correlated positive selection scores.
Table 1.
Pathogen selection across worldwide populations.
Fig 4.
Genes that interact with HIV-1 under selection.
The genes are sorted from left to right by highest to lowest mean selection across the 19 populations exhibiting a p-value < 0.1 for selection associated with HIV-1. The cumulative effect of a moderate selection signal allows for the detection of positive selection in these genes. The top 3 genes are KARS, NGLY1, and POLR2K (leftmost on the axis) are known inhibit HIV-1 replication.
Table 2.
Overrepresentation of HIV human interactome genes in GWASes.
Fig 5.
Worldwide selection of Yersinia pestis.
A p-value for selection in human genes that interact with Yersinia pestis was produced for each population. Each population is represented by a colored circle on the map. The significance for selection is portrayed by a white to red gradient with lighter colors representing higher p-values and brighter red colors representing lower p-values. Only SNPs within genes that interact with Yersinia pestis were included in determining the p-values of selection for each population. These SNPs were filtered for LD. The most significant p-values were found in Europe, Central South Asia, and Africa.
Fig 6.
Overlap of selection targets associated with Yersinia pestis.
Each population was associated with a positive selection score for each gene interacting with Yersinia pestis. There was significant overlap in the genes undergoing selection across both populations in many populations. We performed pairwise correlations among the 11 most selected populations to uncover the extent of overlap in selection targets. Kendall’s’ Tau-b coefficient ranged from -0.09 to 0.22. We identified 23 population pairs with significantly correlated positive selection scores (a positive selection score is produced for each gene).
Fig 7.
Human proteins that interact with Y. Pestis proteins under selection.
The genes are sorted from left to right by highest to lowest mean selection across the 9 populations represented. Only populations with a p-value < 0.1 for selection in proteins that interact with HIV-1 were included in this figure. Selection is primarily driven by moderately selected genes. Genes such as ASH1L, CDIP1, TMEM60, and ZKSCAN5 exhibit strong selection across multiple populations. Genes with the most significant p-values for selection are shown from left to right. Some genes do no exhibit a positive selection score (|iHS|> 2) in any population, indicating that positive selection is detected only by enrichment as no single population exhibits strong selection for such genes.
Table 3.
Y. Pestis selection across worldwide populations.