Three domestications of Asian rice

Domesticated rice (Oryza sativa L.) accompanied the dawn of Asian civilization and has become one of world's staple crops. Despite its global importance and availability of genomic data, questions regarding the phylogenetic, geographic and temporal origin of rice remain unresolved. Although archaeological evidence suggests independent origins of rice cultivation in India and China, paleobotanical findings are relatively scarce and intra-specific classification of rice remains is problematic. On the other hand, genetic studies brought about a variety of contradictory conclusions about the origins of different rice groups, with the most recent studies favouring a single domestication scenario. The difficulties in establishing the true genealogy of the crop stem from the inherent complexity of the domestication process where combined forces of artificial selection, incomplete lineage-sorting and potential migration and inter-group hybridizations obscure the phylogeographic patterns. Here we reanalyzed the largest genetic dataset available for extant wild and cultivated rice, and by comparing the footprints of selection in the genomes of different cultivated rice types, we show that there were three independent domestications in different parts of Asia. We identify wild populations in southern China and the Yangtze valley as the source of japonica gene-pool, and populations in Indochina and the Brahmaputra valley as the source of indica gene-pool. We reveal a hitherto unrecognized origin for the aus variety in central India or Bangladesh. We also conclude that aromatic rice is a result of a hybridization between japonica and aus, and that the tropical and temperate versions of japonica are later adaptations of a single crop. We anticipate that our results will stimulate a more productive collaboration between genetic and archaeological studies of rice domestication, and guide utilization of genetic resources in breeding programmesaimed at crop improvement.