Replicated anthropogenic hybridisations reveal parallel patterns of admixture in marine mussels.

Anthropogenic activity can create new contacts between genetically differentiated lineages, which start to exchange genes again before our eyes. When similar introductions occur in different places, they provide replicated large-scale hybridisation experiments. We studied the population genetics of Mytilus mussels in several M. galloprovincialis introduction areas. Using a dataset of 4279 mussels, genotyped at 77 ancestry informative markers, we document the introduction of the Mediterranean lineage of M. galloprovincialis to five ports in France and of the Atlantic lineage in Norway. In each case, the introduced species has extensively admixed with M. edulis . In Ports, we also observed sharp genetic shifts, at a very fine spatial scale, at the open sea-port interface, where the seaward native population is M. edulis . The sole exception was the bay of Brest, where the native population is the Atlantic lineage of M. galloprovincialis , and where the invading mussels started to spread and admix with the native background. Sharp transitions at the port entrance, at scales below the dispersal distance, do not fit well with the predictions of a migration-selection tension zone model, and instead suggest habitat choice and adaptation to the environment of the port, possibly coupled with connectivity barriers. We then examined departures from the average admixture rate at different loci (i.e., Barton9s concordance). We compared the two human-mediated admixture events (French ports and Norway), to naturally admixed populations, and to transmission ratio distortions observed in experimental lab crosses. When the same M. galloprovincialis background was involved, we found positive correlations in the departures of marker loci in the different locations; but when different M. galloprovincialis backgrounds were involved, we found no or negative correlations. Some of the positive correlations we observed are best explained by a shared history during saltatory colonisation, but some independent replicates imply parallel selective processes, suggesting that the genome-wide effect of admixture is highly repeatable.