Additional file 2 of On the origin and diversification of Podolian cattle breeds: testing scenarios of European colonization using genome-wide SNP data

Additional file 2: Figure S1. All modelled scenarios for colonization tested in the ABC framework. Description: In all the tested scenarios, we assumed that taurine and indicine cattle separated first. Subsequent reduction in effective population size was modelled to take the two independent domestication events that occurred in the Fertile Crescent and the Indus Valley into account. From these known evolutionary events, two sets of scenarios were built. The first three scenarios mirror to two different waves of migration, an early Neolithic migration involving non-Podolian taurine cattle and a secondary migration involving Podolian cattle after their genetic admixture with indicine cattle that occurred in South-Western Asia. On the opposite, the other three scenarios reflect a single taurine diffusion that occurred after the admixture event between indicine and taurine. Within each set of scenarios, we drew different hypotheses of colonization. A Mediterranean route (Scenarios 1 and 4) in which the Italian Podolian breeds mainly derived from the South-Eastern Mediterranean region and introduced via sea trade. To model this scenario, we assumed that from an ancestral population of size N3 located in the South-Eastern Mediterranean region, a first colonization prompted the split between the Balkan and Central Europe breeds, while a subsequent split, driven by sea routes, led to the formation of the Italian Podolian breeds. In this latter separation, we incorporated a reduction in effective population size to accommodate a founder effect as expected when an population introduced by sea trade starts spreading from few individuals. A Balkan route (Scenario 2 and 5) in which we assumed a terrestrial model of colonization, that therefore we modelled from an ancestral population of size N3 located in South-Eastern Mediterranean, a first split gave rise to Balkan and Continental Podolian breeds while a subsequent divergence event led to the origin of the Podolian cattle in Italy. In this model, we do not constrain priors on effective population size in order to simulate a long-term dispersal as a consequence of a terrestrial expansion. Finally, the admixture scenario simulates an equal contribution of the two routes followed by admixture events leading to the formation of the Italian Podolian breeds.