Ceratonia siliqua
For the Treemix analysis, we rooted C. siliqua phylogenetic tree with SM following SVDquartets reconstructions (Fig. 2). Tree topology reconstructed by Treemix (Fig. 3) is highly similar to that of SVDquartets, showing the same strong west-east differentiation. The position of NM was resolved as intermediate between Western (SM, SWS, SES) and Central and Eastern CEUs (CM, NEM, SEM), which agrees with the highly admixed pattern in NM inferred by snmf (Fig. 2 A) and with its low to moderate differentiation with respect to other CEUs (Tab. S2 in supplementary material). The Treemix model without migrations explained 96 % of the covariance (Fig. 3A), and the addition of four migrations resulted in 99.8 % of the total covariance explained (Fig. 3B). Three of the four migrations identified SEM as the source of introgression, into SM, SWS and NM, whereas the fourth migration connected the central-eastern CEUs to SES, again indicating a westward migration.
Impact of carob cultivation effort on carob genetic diversity
Overall genetic diversity values are reported in Table 1. The Nei diversity (Hexp) values decrease from west CEUs and natural or seminatural habitats to east and/or cultivated habitats, regardless of the markers. Interestingly, this trend is negatively correlated with the association index (rbardD), which reports the highest lack of mixing in cultivated CM carobs. The fixation index (f) suggests a lack of heterozygosity in SM, NM, and NEM (and SEM only for SSR), significant excess in CM, but was non significantly different from zero in SWS and SES (and SEM too for RADseq. Based on the differentiation between genetic groups formed by the combination of CEUs and habitats (Table 1), OUTFLANK did not detect candidate loci with a false discovery rate (qvalue) below 0.05, which indicates that the hypothesis of neutrality cannot been rejected (Fig. S6 in supplementary material).