Fig 1.
Geographic pattern of linkage disequilibrium (LD, distance-corrected r2 between pairs of SNPs) in North American Arabidopsis lyrata (A&B) and the correlation of zygosity (Δ) scaled by the level of genome-wide heterozygosity (θ) (C&D).
(A) Average genome-wide LD for intergenic regions. (B) Average LD for the gene FPA that is involved in the autonomous pathway of flowering time regulation. Darker gray shading for A&B indicates higher levels of LD; both surfaces were estimated based on outcrossing populations only. Minimum convex polygon hulls surrounding populations of the western and eastern ancestral genetic clusters are drawn in black. Triangles indicate the core areas from which recolonization began after the most recent glacial maximum. Lines from these cores are recolonization routes revealed by projecting the population phylogeny on the map. Circles are the populations sampled in this study: outcrossing (black), mixed-mating (pink), and selfing (red). Unshaded areas within the polygon hulls are regions with missing data for outcrossing populations. The dashed blue line indicates the maximum extent of the ice sheet during the last glacial maximum. LD increased with increasing distance from the core areas. The correlation of zygosity (Δ) across intergenic (C) and genic (D) regions for 104 individual genomes as estimated by mlRho. Solid lines depict individuals from the eastern genetic cluster and dashed lines from the western genetic cluster, respectively. Individuals from mixed-mating populations are depicted in pink, and from selfing populations in red. For individuals from outcrossing populations the grayscale gradient indicates individuals from the range center (gray) to the range edge (black).
Table 1.
Results of the linear mixed-effects models on genome-wide linkage disequilibrium (LD, distance-corrected r2 between pairs of SNPs) or the correlation of zygosity (Δ) of genic regions and (top) intergenic regions (bottom).
For each fixed effect, the model estimate (β) as well as the chi-square statistics (χ2) with the associated p value from type III Wald χ2 tests are given. For the correlation of zygosity, the results for Δ at 1000 bps distance are shown (see also S7 Fig). Only results for fixed effects are shown.
Fig 2.
Marginal effects for the interactions of LD (r2) estimated from pool-seq between variants that were identified by SIFT to be both deleterious, deleterious and tolerated, or both tolerated, and expansion distance (km, log-scale) (A) and mating system (FIS) (B). Regression lines and their 95% confidence intervals were estimated from a linear mixed-effects model (see Table 2).
Fig 3.
Relationship between Watterson’s theta (θ) (A), mutational load (B) and average genome-wide linkage disequilibrium (LD, distance-corrected r2 between pairs of SNPs). Mutational load was estimated as the product of non-synonymous polymorphic sites (Pn) and their mean derived allele frequency (fn) divided by the analogous product but for synonymous polymorphic sites (Psfs). Each symbol represents the average estimate of a population. Open and filled symbols are the genomic estimates for intergenic and genic regions, respectively. Circles represent populations from the west, triangles those from the east. Symbol colour indicates the mating system: outcrossing (black), mixed-mating (pink), and selfing (red). The model-predicted regression lines are shown for intergenic (dashed line) and genic (solid line) regions.
Table 2.
Results of the linear mixed-effects model on genome-wide linkage disequilibrium (LD) using only distance-corrected r2 between pairs of SNPs in genic regions that were annotated by SIFT.
For each fixed effect, the model estimate (β), as well as the chi-square statistics (χ2) with the associated p value from a type III Wald χ2 tests are given. Only results for fixed effects are shown.
Fig 4.
Relationship between average linkage disequilibrium (LD, distance-corrected r2 between pairs of SNPs) in the two genes flanking the self-incompatibility locus and mating system, expressed by the population inbreeding coefficient (FIS): AL7G32710-ARK3 (A) and AL7G32750-Ubox (B). Each symbol represents the average estimate of a population. The symbol type reflects the genetic clusters (circle for west, triangle for east) and the symbol colour the mating system: outcrossing (black), mixed-mating (pink), and selfing (red). Significance is reported for the effects of expansion distance and mating system, from separate linear models on the data for each gene. For the gene AL7G32750-Ubox in B, the relationship with mating system was not significant following a False Discovery Correction.
Fig 5.
Watterson’s θ across all studied genic regions for each population against expansion distance (or distance from core for rear-edge populations, in log-scale).
Each symbol represents the average estimate of a population. The symbol type reflects the genetic cluster (circle for west, triangle for east) and the symbol colour the mating system: outcrossing (black), mixed-mating (pink), and selfing (red). Bars depict the difference in average genetic diversity between non-outlier genes and outlier genes linked to expansion distance (green) or mating system (blue).