Hidden variation in polyploid wheat drives local adaptation

Abstract

Wheat has been domesticated into a large number of agricultural environments and has the ability to adapt to diverse environments. To understand this process, we survey genotype, repeat content, and DNA methylation across a bread wheat landrace collection representing global genetic diversity. We identify independent variation in methylation, genotype, and transposon copy number. We show that these, so far unexploited, sources of variation have had a significant impact on the wheat genome and that ancestral methylation states become preferentially "hard coded" as single nucleotide polymorphisms (SNPs) via 5-methylcytosine deamination. These mechanisms also drive local adaption, impacting important traits such as heading date and salt tolerance. Methylation and transposon diversity could therefore be used alongside SNP-based markers for breeding.

Figure 1.

Geographical origins combined with hierarchical cluster analysis on 104 accessions from the Watkins core collection plus Chinese Spring wheat. (A) Geographical positions of the accessions color-coded by their allocated cluster from B after SNP hierarchical clustering. (B) Dendrogram constructed using the complete linkage method within the R package hclust to cluster accessions based on SNP allele frequency across 53,341 SNP sites. The tree was cut into eight groups (excluding the reference Chinese Spring) using the R package cutree, and these clusters are color-coded (Methods). (C) Geographical positions of the accessions color-coded by their allocated cluster from D after CpG SMP hierarchical clustering. (D) Dendrogram constructed using the complete linkage method within the R package hclust to cluster accessions based on methylation levels across 18,965 CpG SMP sites (taken from the 359,500 SMPs that were identified). The tree was cut into eight groups using the R package cutree, and these clusters are color-coded (Methods). (E) SNP-based dendrogram from B with individual accessions color-coded as per their cluster from the SMP-based dendrogram from D. For geographical accession positions in A and C, squares outlined in black represent accessions with detailed positional information that is used for plotting; squares with no outline represent accessions with only a country of origin. AU P-values were computed for the main clusters in B and D using the R package pvclust and are shown in red (Methods).

Figure 2.

Visualizing methylation levels for the 105 wheat accessions across 359,500 SMP sites. Using sites with coverage in all 104 Watkins collection accessions plus Chinese Spring, we generated heat maps for methylation levels across (A) CpG-SMPs, (B) CpG-SMPs with accessions ordered by genotype using the heat map from A with accessions reordered based on Figure 1B's SNP clustering dendrogram (shown on top horizontal axis), and (C) non-CpG SMPs. Rows correspond to individual SMP sites and columns indicate accessions. The colored row labels (barcodes) on the left of the heat map indicate which genomic location a SMP falls into (see legend). SMP sites are ordered by their total methylation across the accessions on the vertical axes, and accessions are clustered by SMP profiles on the horizontal axes (Methods).

Figure 3.

Analyzing methylation profiles across the Watkins collection. (A) Violin plots show the percentage of accessions showing methylation per site. Analyzed sites include Tri-genome, Bi-genome, and Uni-genome methylated sites. A comparative subset of 11,769 sites was used for each category. (B) Ancestral methylation associates with an increased SNP rate. The percentage of methylated versus nonmethylated Aegilops tauschii cytosines that show a different allele in Watkins. (C) Ancestral methylation demonstrates that 5-methylcytosines are preferentially deaminated to thymine. The percentage of methylated versus nonmethylated Ae. tauschii cytosines with a C-to-T/G-to-A transition across the Watkins collection. (D) Accession clustering based on the gene families targeted by methylation. Many accessions from the same geographical origin show the same gene families targeted by methylation and are, thus, clustered close to each other in the Accessions axis (vertical dendrogram). Alongside the vertical dendrogram, the two columns of row barcodes (left and right) correspond to the SMP clusters in Figure 1D and SNP clusters in Figure 1B, respectively. (E) Clustering of the 12 accessions subjected to RNA-seq using average gene expression across the replicates for genes showing differential expression between at least two lines (after log₂ transformation). The horizontal dendrogram has its three main clades labeled 1, 2, and 3. Below the horizontal dendrogram, the two barcode rows (top and bottom) correspond to the SMP and SNP clusters in Figure 1, D and B, respectively. Accessions are labeled by line number, country of origin, and phenotype, i.e., TGW (thousand grain weight), HD (heading date), GW (grain width), or Height, with maximum values in green and minimum values in red. Accessions showing the phenotypic tails for heading date that are of interest within this study are highlighted on the horizontal dendrogram with a blue box.

Figure 4.

Analyzing transposable element methylation profiles across the Watkins collection. (A) Base-space per Watkins accession aligned to DNA-transposons and retrotransposons in comparison to Chinese Spring (Methods). (B) Base-space per Watkins accession aligned to DNA-transposons in comparison to Chinese Spring. (C) Base-space per Watkins accession aligned to retrotransposons in comparison to Chinese Spring. (D) Base-space per Watkins accession aligned to retrotransposon-SINE in comparison to Chinese Spring, plotted versus the total cumulative percentages of enriched cytosine residues (gene-associated) that were methylated for CpG, CHG, and CHH methylation. (E) as per D but for retrotransposon-LTR;Gypsy. (F) as per D but for DNA-transposon-TIR;CACTA. (G) Base-space per Watkins accession aligned to retrotransposon-SINE in comparison to Chinese Spring plotted versus the total cumulative percentages of TE-associated cytosine residues that were methylated for CpG, CHG, and CHH methylation. (H) as per G but for retrotransposon-LTR;Gypsy. (I) as per G but for DNA-transposon-TIR;CACTA.