Page 5 - Voda_Dapporto_alii_2016
P. 5

www.nature.com/scientificreports/















































                                Figure 4.  (a) Principal Coordinates Analysis (PCoA) showing the pattern of similarity among islands based
                                on the mean population differentiation (Gst) among the 27 studied species for the 15 areas (Axis 1 and 2). The
                                vertical axis represents the mean Dst of the species on each island; the top right coloured square is the RGB
                                configuration used to attribute colours to the areas. (b) The coloured dots in the bi-dimensional representation
                                of PCoA have been plotted on a map (equirectangular projection) showing the − 100 m isobaths for north
                                Africa and Sicily (black line), which represent a reliable reconstruction of seashores during the LGM. The
                                background map was cropped from a map available at Natural Earth (www.naturalearthdata.com).


                                Supplementary Table S1. The Principal Coordinate Analysis (PCoA) based on the mean pairwise Gst values
                                among areas showed that Europe and north Africa were genetically well-differentiated. Ustica, the Aeolian, the
                                Aegadian and Maltese islands were genetically similar to Sicily-Italian Peninsula, while Lampedusa was more
                                similar to north Africa. Pantelleria and Marettimo had a genetic signal intermediate between the two continents
                                (Fig. 4a,b, Supplementary Note). When the mean genetic diversification among populations, measured as Dst, for
                                the species occupying each area was plotted on the z-axis, islands with intermediate geographic positions were
                                shown to harbour species with low population genetic differentiation (Fig. 4a).
                                Correlations between species occurrence and genetic variation.  As expected, population genetic
                                differentiation of species (measured with Dst) was significantly anti-correlated with dispersal tendency (Spearman
                                rho =  − 0.409, P =  0.028, Fig. 3c). However, Dst was not significantly anti-correlated with species occurrence
                                on islands (Spearman rho =  − 0.051, P =  0.791). The graph representing occurrence on islands (x-axis) and
                                population genetic differentiation (Dst, y-axis) showed a non-random distribution of the species occurring
                                in circum-Sicilian islands: the sector of widespread and genetically diversified species was significantly empty
                                (Fig. 5a). When communities were examined individually, the faunas of Stromboli, Lipari, Salina, Marettimo, and
                                Linosa showed a significant negative correlation between Dst and frequency (Supplementary Table S5). Moreover,
                                the widespread and genetically undiversified species were always more common than expected in a random dis-
                                tribution. On all islands except for Lampedusa, Ustica and Vulcano there was a deficit of rare, genetically homo-
                                geneous species (Supplemetary Table S5 and Fig. 5). The Aeolian Islands, together with Levanzo and the Maltese
                                Islands did not show a significant deficit of rare and genetically diverse species.
                                Discussion
                                Differentiation of populations at an intercontinental crossroad.  The integration of faunistic and genetic
                                differentiation information ultimately allowed the recognition of the main processes shaping island communities in


         Scientific RepoRts | 6:28828 | DOI: 10.1038/srep28828                                                 5
   1   2   3   4   5   6   7   8   9   10