Search

We epigenotyped 211 individuals from 17 Zingiber kawagoii populations using methylation-sensitive amplification polymorphism (MSAP) and investigated the associations of methylated (mMSAP) and unmethylated (uMSAP) loci with 16 environmental variables. Data regarding genetic variation based on amplified fragment length polymorphism (AFLP) were obtained from an earlier study. We found a significant positive correlation between genetic and epigenetic variation. Significantly higher mean mMSAP and uMSAP uHE (unbiased expected heterozygosity: 0.223 and 0.131, respectively, p < 0.001) per locus than that estimated based on AFLP (uHE = 0.104) were found. Genome scans detected 10 mMSAP and 9 uMSAP FST outliers associated with various environmental variables. A significant linear fit for 11 and 12 environmental variables with outlier mMSAP and uMSAP ordination, respectively, generated using full model redundancy analysis (RDA) was found. When conditioned on geography, partial RDA revealed that five and six environmental variables, respectively, were the most important variables influencing outlier mMSAP and uMSAP variation. We found higher genetic (average FST = 0.298) than epigenetic (mMSAP and uMSAP average FST = 0.044 and 0.106, respectively) differentiation and higher genetic isolation-by-distance (IBD) than epigenetic IBD. Strong epigenetic isolation-by-environment (IBE) was found, particularly based on the outlier data, controlling either for geography (mMSAP and uMSAP βE = 0.128 and 0.132, respectively, p = 0.001) or for genetic structure (mMSAP and uMSAP βE = 0.105 and 0.136, respectively, p = 0.001). Our results suggest that epigenetic variants can be substrates for natural selection linked to environmental variables and complement genetic changes in the adaptive evolution of Z. kawagoii populations.

Ecological and evolutionary processes linking adaptation to environment are related to species’ range shifts. In this study, we employed amplified-fragment-length-polymorphism-based genome scan methods to identify candidate loci among Zingiber kawagoii populations inhabiting varying environments distributed at low to middle elevations (143–1488 m) in a narrow latitudinal range (between 21.90 and 25.30° N). Here, we show evidence of selection driving the divergence of Z. kawagoii. Twenty-six FST outliers were detected, which were significantly correlated with various environmental variables. The allele frequencies of nine FST outliers were either positively or negatively correlated with the population mean FST. Using several independent approaches, we found environmental variables act in a combinatorial fashion, best explaining outlier genetic variation. Nonetheless, we found that adaptive divergence was affected mostly by annual temperature range, and it is significantly positively correlated with latitude and significantly negatively correlated with the population mean FST. This study addresses a latitudinal pattern of changes in annual temperature range (which ranged from 13.8 °C in the Lanyu population to 18.5 °C in the Wulai population) and emphasizes the pattern of latitudinal population divergence closely linked to the allele frequencies of adaptive loci, acting in a narrow latitudinal range. Our results also indicate environmentally dependent local adaptation for both leading- and trailing-edge populations.

The testing association of environmental variables with genetic and epigenetic variation could be crucial to deciphering the effects of environmental factors playing roles as selective drivers in ecological speciation. Although ecological speciation may occur in closely related species, species boundaries may not be established over a short evolutionary timescale. Here, we investigated the genetic and epigenetic variations using amplified fragment length polymorphism (AFLP) and methylation-sensitive amplification polymorphism (MSAP), respectively, and tested their associations with environmental variables in populations of four closely related species in the R. pseudochrysanthum complex. No distinctive species relationships were found using genetic clustering analyses, neighbor-joining tree, and neighbor-net tree based on the total AFLP variation, which is suggestive of the incomplete lineage sorting of ancestral variation. Nonetheless, strong isolation-by-environment and adaptive divergence were revealed, despite the significant isolation-by-distance. Annual mean temperature, elevation, normalized difference vegetation index, and annual total potential evapotranspiration were found to be the most important environmental variables explaining outlier genetic and epigenetic variations. Our results suggest that the four closely related species of the R. pseudochrysanthum complex share the polymorphism of their ancestor, but reproductive isolation due to ecological speciation can occur if local environmental divergence persists over time.