Machine learning and phylogenetic models identify predictors of genetic variation in Neotropical amphibians

Abstract

Many factors influence genetic variation within species, which is key for adaptation and survival in changing environments. Using amphibians as a study system, we investigated genetic diversity and structure within Neotropical species to provide insights into how natural history traits, phylogeny, and geographic characteristics influence intraspecific genetic variation. We assembled datasets for Neotropical amphibians using open-access databases for natural history traits, genetic sequences, phylogenetic trees, and geographic data. For each species, we calculated overall nucleotide diversity (π) and tested for isolation by distance (IBD). We then identified predictors of genetic diversity and genetic structure (IBD) using Random Forest (RF) regression or RF classification, respectively. To more explicitly incorporate phylogenetic relationships, we also used phylogenetic generalized linear mixed models (PGLMMs), and phylogenetic linear regression or phylogenetic logistic regression models to predict π and IBD, respectively. We compiled > 4,000 mitochondrial DNA sequences from 256 amphibian species (230 frogs and 26 salamanders), georeferencing 2,468 sequences from 176 species that were not already linked to occurrence data. RF regressions, phylogenetic linear regression, and PGLMMs were congruent in identifying range size and elevation as the most important predictors of π. Species with larger ranges and at lower elevations tended to have higher overall genetic diversity. RF classification and phylogenetic comparative methods identified minimum elevation as the most important predictor of IBD, with species at lower elevations being more likely to exhibit genetic structure. These results could play a key role in the development of conservation strategies for amphibians. For example, species living at high elevations or with small ranges need substantial protection due to their low genetic diversity. We also show that the Choco, Andes, and Atlantic Forests biogeographic units harbor high diversity across many species that should be prioritized for protection.

Predicting genetic biodiversity in salamanders using geographic, climatic, and life history traits

“…Similarly, in amphibians, several recent studies have found within species genetic diversity to be lower in temperate regions in species with smaller ranges and at higher elevations (Barrow et al, 2020;Amador 2023). The methods used to detect these patterns are based on current taxonomic knowledge, and as such, rely on the assumption that the species designations used are accurate.…”

“…This suggests that the presence of variation in climate, rather than any species-specific trait or characteristic is the most identifiable driving force of within species genetic diversity for salamanders at this scale. Species traits were not a predictor of intraspecific genetic diversity in amphibians (Barrow et al, 2021;Amador et al, 2023) using a different measure of genetic variation within species (nucleotide diversity). Using similar methods, our results in salamanders differ from that found in mammals, where body size and range size were the most important predictors (Parsons et al, 2022).…”

“…Conversely, in other organisms, isolationby-distance within species is often higher at higher latitudes (multiple taxonomic groups: amphibians: Amador et al, 2023). Further, genetic variation within amphibians was best explained by range size and elevation, rather than latitude, in the neotropics (Amador et al, 2023), while latitude was an important predictor of genetic diversity in the nearctic (Barrow et al, 2021). This suggests that differences exist in how genetic variation is distributed within species depending on which taxonomic groups are being examined, and at what spatial scale.…”