Genetic architecture, demographic history, and genomic differentiation of Populus davidiana revealed by whole-genome resequencing

Abstract

Forest trees are an excellent resource from which to understand population differentiation and heterogeneous genome variation patterns due to the majority of forest trees being distributed widely and able to adapt to different climates and environments. Populus davidiana is among the most geographically widespread and ecologically important tree species in China. Whole-genome resequencing data of 75 individual examples of P. davidiana throughout China were conducted, finding that all examples from different regions were clearly divided into either Northeast (N), Central (C), and South (S) populations. The ancestors of P. davidiana diverged into Northern group, comprising both N and C and Southern populations approximately 792,548 years ago. This time point of differentiation suggests that divergence of P. davidiana populations might have been triggered by the mid-Pleistocene transition. The three populations experienced considerable periods of bottleneck following divergence, with population expansion beginning around 5,000 years ago after the end of the last glacial maximum. We found N to be the center of origin of P. davidiana in China. The migration route of P. davidiana in China was from N to S. Although the majority of the regions of genomic differentiation between N and S populations can be explained by neutral processes, a number of tested outlier regions were also found to have been significantly influenced by natural selection. Our results highlight that linked selection and rates of recombination were important factors in genomic differentiation between the N and S populations. Finally, we identified a substantial number of functional genes related to climate change during population differentiation and adaptive evolution.

Keywords: Populus davidiana; demographic history; genetic architecture; heterogeneous genomic differentiation; whole‐genome sequencing.

Figure 1

Geographical distribution of 75 Populus davidiana

Figure 2

Principal component analysis (PCA) was performed using PCAngsd

Figure 3

(a) Best‐fitting model inferred demographic histories and differentiation mode for three different P. davidiana populations implemented by fastsimcoal 2.6.1. (b) The effective population size (Ne) over historical time implementing by MSMC

Figure 4

Pairwise F _ST between P. davidiana populations through each chromosome

Figure 5

Population genetic analysis. (a) Pairwise F _ST between P. davidiana populations through the whole genome; (b) the difference between two nucleotides (dxy); (c) nucleotide diversity (θ_π); (d) Tajima's D; (e) Fay & Wu's H; (f) Watterson's estimator (θ_W); (g) gene diversity/heterozygosity (HE); (h) recombination rate (ρ)

Figure 6

The outlier regions that have been tested to be significantly influenced by natural selection. (a) Comparisons of dxy among regions displaying significantly high (blue boxes) and low (red boxes) differentiation versus the genomic background (green boxes) between N and S populations; (b) comparisons of RND between N and S populations; (c) comparisons of nucleotide diversity π in N; (d) comparisons of nucleotide diversity π in S; (e) comparisons of Tajima's D in N; (f) comparisons of Tajima's D in S; (g) comparisons of Fay & Wu's H in N; (h) comparisons of Fay & Wu's H in S; (i) comparisons of r ² in N; (j) comparisons of r ² in S; (k) comparisons of recombination rate (ρ/θ_π) in N; (l) comparisons of recombination rate (ρ) in S. Asterisks designate significant differences between outlier windows and the rest of genomic regions by Mann–Whitney U test (*p‐value < .05; **p‐value < 1e−4; ***p‐value < 2.2e−16)

Figure 7

The proportion of interspecific shared polymorphisms (a) and the proportion of fixed differences caused by derived alleles fixed in either N (b) or S (c)