Population Genomics and Lagrangian Modeling Shed Light on Dispersal Events in the Mediterranean Endemic Ericaria zosteroides (=Cystoseira zosteroides) (Fucales)

Dispersal is a central process that affects population growth, gene flow, and ultimately species persistence. Here we investigate the extent to which gene flow occurs between fragmented populations of the deep-water brown algae Ericaria zosteroides (Turner) Greville (Sargassaceae, Fucales). These investigations were performed at different spatial scales from the bay of Marseille (western Provence) to Corsica. As dispersal of zygotes is shown to be limited over distances beyond a few meters, we used a multidisciplinary approach, based on Lagrangian modeling and population genomics to test the hypothesis that drifting of fertile parts of thallus (eggs on fertile branches), mediated by ocean currents, enable occasional gene flow between populations. Therefore we assessed the respective contribution of oceanographic connectivity, geographical isolation, and seawater temperatures to the genetic structure of this species. The genetic structure was assessed using 10,755 neutral SNPs and 12 outlier SNPs genotyped by dd-RAD sequencing in 261 individuals of E. zosteroides. We find that oceanographic connectivity is the best predictor of genetic structure, while differentiation in outlier SNPs can be explained by the depth of populations, as emphasized by the minimum seawater temperature predictor. However, further investigations will be necessary for clarifying how depth drives adaptive genetic differentiation in E. zosteroides. Our analyses revealed that local hydrodynamic conditions are correlated with the very high divergence of one population in the Bay of Marseille. Overall, the levels of gene flow mediated by drifting were certainly not sufficient to counteract differentiation by local genetic drift, but enough to allow colonization several kilometers away. This study stresses the need to consider secondary dispersal mechanisms of presumed low dispersal marine species to improve inference of population connectivity.