Non-linear genetic isolation by distance: implications for dispersal estimation in anadromous and marine fish populations

Abstract

Indirect genetic approaches such as those based on the association between genetic and geographic distance (isolation by distance, IBD) may provide one of the best means of estimating dispersal in marine systems. We evaluated the scale -dependency and the ecological covariates of the 1131) relationship through a combined modeling and meta-analytical approach. Using Wright's fixation index F-ST, simulations were used to generate IBD relationships, F-ST/(1-F-ST) and geographic distance, for various combinations of average dispersal distance and number of generations. 1131) linearity increased with time since colonization and dispersal distance, although the effect of dispersal distance predominated after initial colonization. Simulations suggest that declines in the 1131) slope were associated with increases in the spatial scale of observation, suggesting that the IBD pattern is non-linear at very short and long average dispersal distances. We hypothesized that non-linear IBD would be common in marine populations, and apparent through a biologically significant non-zero intercept arid decreasing slope with increasing geographic scale. Predictions of common non-linearity were examined through IBD relationships (i.e. slope, intercept, R-2 of the regression) and life history parameters from the published literature for 18 species of anadromous/philopatric fishes. As predicted, 1131) parameters (intercept and slope) were consistently correlated with many life history traits (e.g. fecundity, egg size). However, the statistical removal of adult size eliminated most significant life history-IBD, correlations. Increases in IBD slope were associated with decreasing gene flow (p = 0.014, R-2 = 0.33), and decreasing migration distance (p = 0.039, R-2 = 0.23). Non-linearity was further supported by consistent declines in the 1131) slope with increased geographic scale in anadromous fish as well as several marine species, suggesting the ubiquity of this phenomenon. We conclude that isolation by distance patterns may reflect dispersal phenotype and are biologically significant. Nonetheless, non-linearity in 1131) pattern is probably the norm in aquatic organisms, resulting from large ranges and limited dispersal. Accordingly, approximations of demographic parameters based on the IBD must be made cautiously, taking into account possible non-linearity, scale dependencies, and assumptions of genetic drift-dispersal equilibrium.