5. Comparison of eDNA- and tissue-based genetic differentiation
Correlations between the metrics calculated by eDNA- and tissue-based approaches were examined for each statistic of genetic diversity and differentiation. For genetic differentiation, statistics were calculated in the three datasets: the entire dataset, the dataset upstream of the Kanayama dam (upstream dataset), and the dataset only between population pairs with a waterway distance of <15 km (15-km dataset). The upstream dataset was used to exclude the effects of a long spatial gap and a dam, and the 15-km dataset was used to reveal the pattern at the spatial scale where gene flow is especially dominant (Nakajima et al. 2021). Correlations of genetic differentiations for the entire dataset and upstream dataset were tested by Mantel tests (9999 permutations and 10000 times bootstrapping to estimate confidence intervals) using the package ecodist in R, and the statistics for datasets not in matrix style (genetic differentiation in the 15-km dataset and genetic diversity) were tested by simple bootstrapping (10000 times) using the package CarletonStats 2.2 (Chihara and Loy 2023).