Environmental Association Analyses Identify Candidates for Abiotic Stress Tolerance in Glycine soja, the Wild Progenitor of Cultivated Soybeans

Abstract

Natural populations across a species range demonstrate population structure owing to neutral processes such as localized origins of mutations and migration limitations. Selection also acts on a subset of loci, contributing to local adaptation. An understanding of the genetic basis of adaptation to local environmental conditions is a fundamental goal in basic biological research. When applied to crop wild relatives, this same research provides the opportunity to identify adaptive genetic variation that may be used to breed for crops better adapted to novel or changing environments. The present study explores an ex situ conservation collection, the USDA germplasm collection, genotyped at 32,416 SNPs to identify population structure and test for associations with bioclimatic and biophysical variables in Glycine soja, the wild progenitor of Glycine max (soybean). Candidate loci were detected that putatively contribute to adaptation to abiotic stresses. The identification of potentially adaptive variants in this ex situ collection may permit a more targeted use of germplasm collections.

Keywords: Glycine soja; crop wild relative; germplasm collections; landscape genomics; population structure; soybean.

Figure 1

Results of STRUCTURE analysis in G. soja accessions, and the geographical location in which each were collected. The spot colors correspond to the STRUCTURE assignment of each accession, Green: Mainland South; Blue: Mainland North; Red: Island. The assignment of samples into three genetic clusters generally accords with geography. The spots have been jittered to show overlapping samples.

Figure 2

Genome-wide associations with Mean Temperature Wettest Quarter. (A) Manhattan plot of negative log p-values. (B) Zoom in on 60 kb region around the significant markers BARC_1.01_Gm08_40882335_A_G and BARC_1.01_Gm08_40883682_C_T. The Arabidopsis ortholog for a nearby gene, Glyma.08g298200, is MYB88, a gene-associated stomata development. (C) The frequency of nonreference “G” and “T” alleles is high in G. soja, and rare in a previous study of landrace and elite lines. (Note: The marker name is based on the genome assembly version 1 position (e.g., 40882335) while the position shown in (B) is based on the assembly version 2 position (e.g., approximately 41520000). This explanation also applies to the discrepancy between marker name and genome position in Figures 3 and 4.)

Figure 3

Genome-wide association results of percent sand, and percent silt. (A) Genome-wide view of association results for Percent Sand Subsoil. (B) Zoom in on 60 kb region around the significant marker BARC_1.01_Gm14_23750665_G_A, the most significant hit for topsoil and subsoil percent sand, and topsoil and subsoil percent silt. The “A” allele at this locus is associated with high silt environments, and is not found in a previous scan of landrace and elite soybean cultivars. The Arabidopsis ortholog for the nearest gene, Glyma.14g141200, is YUC6, a gene associated with enhanced resistance to water stress. (C) The “A” allele is rare in our sample, and found to be rare, or not present, in a previous screen of soybean genotypic classes (Song et al. 2013). (D) Density plot of allele frequency distribution for Percent Silt. The individuals with the “G” allele are shaded in dark gray overlaid with the “A” allele individuals in light gray. (E) Geographic location of individuals with the “G” allele (dark gray), or “A” allele (light gray), with jitter added to show overlapping samples. Individuals with missing genotyping data at this SNP are not shown.

Figure 4

SPA, F_ST, and recombination rate in the G. soja genome. (A) Sliding window of these values plotted on chromosome 15. Recombination decreases dramatically through the pericentromeric region, denoted by the vertical gray dotted lines. (B) Zoom in on 60 kb region around two significant SPA markers BARC_1.01_Gm15_10376148_G_A, and BARC_1.01_Gm15_10382285_T_C, a region of notably low recombination, and both high F_ST and SPA values. Three genes in this region (denoted with asterisks) were previously found to be duplicated, or deleted, in some elite soybean lines (Anderson et al. 2014). This cluster of genes appear to be members of a gene family. The Arabidopsis ortholog for the genes denoted in red is AT5G46890, a bifunctional inhibitor/lipid-transfer protein/seed storage 2S albumin superfamily protein. Similarly, the Arabidopsis ortholog for Glyma.15g119600, denoted in blue, is AT5G46900, a bifunctional inhibitor/lipid-transfer protein/seed storage 2S albumin superfamily protein. The implications of structural variation relating to F_ST, SPA hits, or recombination are not yet clear.