Role of genetic introgression during the evolution of cultivated rice (Oryza sativa L.)

Abstract Background Models for the origins of cultivated rice currently fall into two groups: ones that identify independent domestications of the indica, japonica and possibly also the aus types, and others that propose that the domestication phenotype was initially acquired by japonica, the underlying alleles then transferred by introgression to other pre-domesticated populations, giving the indica and aus varieties. Identifying the impact of past gene flow on cultivated rice genomes is therefore crucial to distinguishing between these models and understanding the domestication history of rice. To this end, we used population-scale polymorphism data to identify the progenitor gene pools of indica, japonica and aus. Variation shared among the cultivated groups but absent from at least one progenitor population was identified, and genomic blocks putatively transferred by gene flow among cultivated groups mapped. Results Introgression signals were absent at the major domestication loci (Prog1, Rc, qSH1, qSH3, Sh4) of indica and aus, indicating that these loci were unaffected by gene flow from japonica. Other domestication-related loci (Ghd7, LABA1, Kala4, LG1) show signals of introgression from japonica or indica to aus. There is a strong signal for LABA1 in japonica, possibly indicating introgression from indica. The indica genome is the least affected by gene flow, with just a few short regions with allelic frequencies slightly altered by introgression from japonica. Conclusion Introgression has occurred during the evolution of cultivated rice, but was not responsible for transfer of the key domestication alleles between the cultivated groups. The results are therefore consistent with models in which japonica, indica and aus were domesticated independently, with each of these cultivated groups acquiring the domestication alleles from standing variation in wild rice, without a significant contribution from inter-group gene flow.