Deep Analysis of 125 Cultivated and Wild Rice Genomes: Distinguishing Recent Gene Flow from Common Ancestry

Resequencing a population sample of homozygous lines to 10X coverage or higher permits robust polymorphism discovery with nearly complete genotype data across the sample. Initially as part of an international effort (the Rice SNP Consortium) to identify SNPs for the development of robust genotyping assays, 88 cultivated lines representing all five major subpopulations of Oryza sativa and 37 Oryza rufipogon/nivara (wild) accessions were sequenced along with 3 outgroup species, yielding a dataset with an unexpected high level of SNP diversity.  With broad representation of each cultivated subpopulation, we investigated SNP diversity within and between subpopulations as well as common and differentiated haplotypes. With new methods we are able to analyze subpopulation structure at high resolution and decompose the genome of an individual into a mosaic of ancient and recent introgressions. Comparing wild and cultivated populations we ask how a potentially structured ancestral wild population with local ecological adaptations gave rise to five readily distinguished cultivated populations as opposed to a single domesticated population followed by an extensive dissemination to other emerging civilizations. Key to our understanding of early rice domestication is how and where in the genome local ecological adaptations were maintained, and what advantageous alleles for cultivation were transferred between subpopulations. Further analysis of nearly complete genotype datasets without ascertainment or imputation bias may reveal candidate regions impacting breeding efforts for climate change as well as reveal favorable alleles currently confined to one subpopulation which genetic studies and breeding programs have not yet identified.