Signatures of local adaptation to current and future climate in phenology-related genes in natural populations of Quercus robur

Abstract

Background: Local adaptation is a key evolutionary process that enhances the growth of plants in their native habitat compared to non-native habitats, resulting in patterns of adaptive genetic variation across the entire geographic range of the species. The study of population adaptation to local environments and predicting their response to future climate change is important because of climate change.

Results: Here, we explored the genetic diversity of candidate genes associated with bud burst in pedunculate oak individuals sampled from 6 populations in Poland. Single nucleotide polymorphism (SNP) diversity was assessed in 720 candidate genes using the sequence capture technique, yielding 18,799 SNPs. Using landscape genomic approaches, we identified 8 F_ST outliers and 781 unique SNPs in 389 genes associated with geography, climate, and phenotypic variables (individual/family spring and autumn phenology, family diameter at breast height (DBH), height, and survival) that are potentially involved in local adaptation. Then, using a nonlinear multivariate model, Gradient Forests, we identified vulnerable areas of the pedunculate oak distribution in Poland that are at risk from climate change.

Conclusions: The model revealed that pedunculate oak populations in the eastern part of the analyzed geographical region are the most sensitive to climate change. Our results might offer an initial evaluation of a potential management strategy for preserving the genetic diversity of pedunculate oak.

Keywords: Bud-burst phenology; Candidate genes; Forest tree; Genotype-environment association; Local adaptation; Sequence capture.

Fig. 1

Genetic structure of Q. robur in Poland. The pie charts show the proportions of membership of each population for the estimated value of K = 4 genetic groups

Fig. 2

Results for the differentiation ( F_ST) outlier test based on four gene pools of Q. robur in Poland. SNPs with log₁₀q less than -1.3 (corresponding to q-value < 0.05) are considered outliers

Fig. 3

The relative importance (R²) of the predictor variables used in gradient forests (GF) for the four SNP sets

Fig. 4

Projected spatial turnover of Q. robur allele frequencies using GFs for different subsets of SNPs: all SNPs (e); SNPs associated with geography (f); climate (g); and phenotypic traits (h). The biplots (a-d) indicate the contribution of the environmental variables to the predicted patterns of genetic turnover (e–h), with labeled vectors indicating the direction and magnitude of environmental gradients. The difference between GF models (i-k) mapped in (e) and (f–h) is based on Procrustes residuals, scaled by the maximum distance found for each comparison. Black circles (e-k) indicate the locations of genotyped populations

Fig. 5

Predicted genetic offset for the complete dataset of SNPs (a) and the subset of SNPs related to geography (b), climate (c), and phenotypic traits (d) under the climate change scenario for 2080. Black circles (a-d) indicate the locations of genotyped populations