Berry skin colour, and indirectly the anthocyanin content, is probably one of the agronomical traits most easily selected by human during the history of the grapevine. Consequently the anthocyanin biosynthesis pathway has very interesting features for the study of gene network evolution. Moreover, the structural genes involved in this pathway are identified and characterized and regulatory genes (VvMybAs) of this pathway are also known. Our recent data suggest that the genetic determinism of the total variation of anthocyanin content, and by extension the berry skin colour, in cultivated grapevine is rather due to few mutations within the sequences of the regulatory VvMybAs genes than mutations in anthocyanin structural genes. We have sequenced the VvMybAs genes on a core-collection of cultivated grapevine in order to understand the evolution of the genes at the allelic level and to identify the order of emergence of the principal mutations linked with the anthocyanin content. Secondly, we have sequenced the anthocyanin structural genes on another sample consisting of Vitis Vinifera L. (cultivated and wild) and other species of Vitis to identify the direct and indirect impact of human selection in these genes. These studies showed that berry color-associated mutations evolved from wild black alleles to a pink allele (SNP variation in coding sequence of VvMybA2) and to the white one (Insertion of Gret1 in the promoter of VvMybA1). We have also identified a significant link between the positions of genes in the anthocyanin pathway and their level of divergence. More surprisingly, comparison between values of Tajima’s D and Fay and Wu’s H of the wild and cultivated samples for each gene showed a different impact of the selection pattern and different selected targets within these two samples. This analysis is an interesting example for developing this kind of study at a whole genome level.