Restriction associated DNA-genotyping at multiple spatial scales in Arabidopsis lyrata reveals signatures of pathogen-mediated selection

James Buckley,Philippine Vergeer,Barbara K Mable,Marcus A Koch,Eric B Holub

doi:10.1186/s12864-018-4806-7

Abstract

BackgroundGenome scans based on outlier analyses have revolutionized detection of genes involved in adaptive processes, but reports of some forms of selection, such as balancing selection, are still limited. It is unclear whether high throughput genotyping approaches for identification of single nucleotide polymorphisms have sufficient power to detect modes of selection expected to result in reduced genetic differentiation among populations. In this study, we used Arabidopsis lyrata to investigate whether signatures of balancing selection can be detected based on genomic smoothing of Restriction Associated DNA sequencing (RAD-seq) data. We compared how different sampling approaches (both within and between subspecies) and different background levels of polymorphism (inbreeding or outcrossing populations) affected the ability to detect genomic regions showing key signatures of balancing selection, specifically elevated polymorphism, reduced differentiation and shifts towards intermediate allele frequencies. We then tested whether candidate genes associated with disease resistance (R-gene analogs) were detected more frequently in these regions compared to other regions of the genome.ResultsWe found that genomic regions showing elevated polymorphism contained a significantly higher density of R-gene analogs predicted to be under pathogen-mediated selection than regions of non-elevated polymorphism, and that many of these also showed evidence for an intermediate site-frequency spectrum based on Tajima’s D. However, we found few genomic regions that showed both elevated polymorphism and reduced FST among populations, despite strong background levels of genetic differentiation among populations. This suggests either insufficient power to detect the reduced population structure predicted for genes under balancing selection using sparsely distributed RAD markers, or that other forms of diversifying selection are more common for the R-gene analogs tested.ConclusionsGenome scans based on a small number of individuals sampled from a wide range of populations were sufficient to confirm the relative scarcity of signatures of balancing selection across the genome, but also identified new potential disease resistance candidates within genomic regions showing signatures of balancing selection that would be strong candidates for further sequencing efforts.

Highlights

Genome scans based on outlier analyses have revolutionized detection of genes involved in adaptive processes, but reports of some forms of selection, such as balancing selection, are still limited
We identified genomic regions of elevated polymorphism containing disease resistance candidates known to be influenced by balancing selection in A. thaliana
Within regions showing classical signatures of balancing selection, we identified novel candidate disease resistance genes, which have not been previously described as balanced polymorphisms in other taxa

Summary

Introduction

Genome scans based on outlier analyses have revolutionized detection of genes involved in adaptive processes, but reports of some forms of selection, such as balancing selection, are still limited It is unclear whether high throughput genotyping approaches for identification of single nucleotide polymorphisms have sufficient power to detect modes of selection expected to result in reduced genetic differentiation among populations. High throughput sequencing and associated genotyping methods have revolutionized population genomic studies of population structure and selection and been applied to a wide range of species [1] Such approaches have proved useful for resolving phylogeographic structure in non-model taxa [2,3,4], and for detecting signatures of selection across the genome and identifying novel candidate loci being influenced by selection [1, 4,5,6,7,8,9]. An important question is whether this scarcity of evidence is due to balancing selection being limited in its role in shaping genome-wide patterns of diversity (e.g. [13]), or due to the challenges of identifying signatures of balancing selection as opposed to other types of selection [10]

Methods

Results

Discussion

Conclusion