Genetic diversity and tests of selection in PgTb1 and STS control loci.

aThe same collection was sequenced for the STS loci and PgTb1. When some STS loci could not be amplified in a given individual (for technical reasons), they were sequenced in an individual from a geographically close population (details in Table S2), in order to keep sample size and composition consistent with PgTb1 and between STS. Individuals Tb70, Tb83 and Tb84 that are issued from seeds collected in wild populations (see Table S2) displayed a weedy-like phenotype at maturity. For this reason, we preferred to discard their sequences from the wild sample for the genetic diversity analysis. Conclusions from neutrality tests were unchanged when these 3 plants were included (data not shown).bS: number of segregating sites.cSpecific sites: sites polymorphic in one population (e.g. wild) but monomorphic in the other (e.g. domesticated).dθ: Watterson estimator of sequence diversity, π: average number of differences per site between two sequences. For both estimators, the diversity in the PgTb1 mRNA region is broken down into non-coding (nc), synonymous (syn) and non-synonymous (nsyn) sites.‡For the domesticated sample, a p-value of D and F* was calculated for each one of the 54 simulated scenarios of demographic history. This table gives the range of these p-values excepting scenario 39 which was discarded on the basis of its extremely low likelihood (see methods and Figure S3C): significant p-values consistent across all demographic scenarios provide strong support for selection. Refer to Table S3 for a comprehensive list of these p-values per scenario and per region (details of the scenarios and simulation methods in Figure S3).**STS 713 is on linkage group 6, STS 476 on linkage group 3 and STS 738 on linkage group 2 at 27 cM from PgTb1 (referred to as marker Xpsm738 on the genetic map of Figure 4B).