Next-Generation RAD Sequencing Unravel Genomic Diversity and Patterns of Introgression Between Wild and Domesticated Manihots

Cassava (Manihot esculenta Crantz) known as yuca, mandioca, macacheira, manioc or tapioca is the fourth most important source of carbohydrates in the tropics after rice, sugarcane and maize, and the sixth most important crop globally.  This perennial woody shrub was domesticated between 5000 and 7000 BC in the American Tropics, but the exact area in which this took place remains undetermined.  Although, there is a high degree of controversy on the origin, evolution and domestication of cassava, it is widely accepted that cassava was domesticated from M. e. flabellifolia, which is predominantly distributed in Northeastern Brazil, but not exclusively.  This view  is supported by the abundance of wild forms of the genus Manihot in Northeast Brazil. Insights into the crop’s history will be critical to target useful genetic diversity for the crop’s improvement. Molecular phylogenetic/phylogeographic assessments undertaken to unravel the origin of cassava, has given valuable insight into the crop evolution and domestication; however, the phylogenetic relationship among M.e. flabellifolia indicates that those from the French Guyana appear to be not as close to cassava as those from Amazon basin raising several issues into the crops origin and domestication. In cassava, the study of genomic admixture offers a prime opportunity to determine whether or not anthropogenic activities resulted in the movement of M. e. flabellifolia populations, explaining the current genetic separation between M. e. flabellifolia found in French Guyana with those in South West Amazonia and ultimately give insight into the domestication of cassava. This became possible only recently due to the development of: i) an inexpensive approach to discover tens if not hundreds of thousands of SNP using restriction restriction-site-associated DNA (RAD) on the Illumina sequencing platform, ii) novel analytical tools suited to large numbers of codominant markers to detect population structure and genetic admixtures by a Bayesian approach and iii).  A draft of the cassava genome has been generated against which align the RAD sequence reads  confidently identify SNPs on the basis of observed patterns of heterozygosity without the need of further SNP valiadation. We generated on average 703Mbp of RAD sequencing data for each accession and we applied a bioinformatics pipeline to obtain a database of over 500 thousand variable sites, from which close to 200 thousand were variable within the M. esculenta specie. We performed a genome wide population genetics analysis to identify the genetic relatedness among wild and domesticated forms in the genus Manihot.