Detection of adaptive divergence in populations of the stream mayfly Ephemera strigata with machine learning

Adaptive divergence is a key mechanism shaping the genetic variation of natural populations. A central question linking ecology with evolutionary biology concerns the role of environmental heterogeneity in determining adaptive divergence among local populations within a species. In this study, we examined adaptive the divergence among populations of the stream mayfly Ephemera strigata in the Natori River Basin in northeastern Japan. We used a genome scanning approach to detect candidate loci under selection and then applied a machine learning method (i.e. Random Forest) and traditional distance-based redundancy analysis (dbRDA) to examine relationships between environmental factors and adaptive divergence at non-neutral loci. We also assessed spatial autocorrelation at neutral loci to quantify the dispersal ability of E. strigata. Our main findings were as follows: 1) random forest shows a higher resolution than traditional statistical analysis for detecting adaptive divergence; 2) separating markers into neutral and non-neutral loci provides insights into genetic diversity, local adaptation and dispersal ability and 3) E. strigata shows altitudinal adaptive divergence among the populations in the Natori River Basin.