Inferring gene flow in coral reef fishes from different molecular markers: which loci to trust?

Contrasting results are usually reported in the literature regarding the factors influencing observed structuring of genetic variability. The goals of this study were, for five coral reef fishes in French Polynesia, (1) to infer the theoretical variance of single locus F(ST) estimates expected under neutrality in order to exclude outlier loci before inferring gene flow and (2) to test thereafter whether species laying pelagic eggs effectively disperse more than species laying benthic eggs in this system. For this purpose, a total of 952 individuals from five species belonging to two families (Chaetodontidae and Pomacentridae) were screened among populations sampled within a 60-600 km spatial range for intron length polymorphism at 11 loci in order to illuminate contrasting results previously published on allozymes and mitochondrial DNA (mtDNA) control region polymorphisms. Statistically speaking, among the five species, four loci (three allozymes and one intron) were identified as outliers and discarded before interpretation of genetic differentiation in terms of effective dispersal. Biologically speaking, our results suggest that the observed genetic structure is not significantly related to the reproductive strategy of coral reef fish in the island system we analysed and that observed random genetic differentiation accommodates Wright's island model in all five species surveyed. Overall, our study emphasizes how cautious one has to be when trying to interpret present-day genetic structure in terms of gene flow while using a limited number of loci and/or different sets of loci.