Genome-wide diversity of coconut from northern South America uncoversgenotypes present in Colombia and strong population structure

Abstract

Coconut palms (Cocos nucifera) are a combination of wild admixed populations and perennial crops with a worldwide distribution. Here we develop single nucleotide polymorphisms (SNPs) along the coconut genome based on Genotyping by Sequencing (GBS) for at least four different commercially important and widely cultivated coconut varieties and hybrids growing in northern South America. We present a comprehensive catalog of approximately 27K SNPs to conduct genetic diversity, population structure and linkage disequilibrium analysis. A relatively fast LD decay for the Atlantic accessions within ~250Kb was observed in comparison to the Pacific accessions ~ 1500 Kb. The complete SNPs sampling showed a strong population structure at K = 2, separating accessions from the Pacific and Atlantic coasts as it has been found in previous studies. At higher K values, one non-admixed group was observed for the Atlantic while further substructures emerged for the Pacific accessions, where three non-admixed groups were found. Population structure analysis also showed a great degree of admixture between the Atlantic and Pacific populations, and SNPs of the Pacific non-admixed genetic groups were mostly introgressed into the Atlantic individuals but the contrary was rarely observed. The results of principal component analysis and Neighbor-Joining Hierarchical Clustering were consistent with the results from Structure and provided a measure of genetic relationships among individual genotypes. The Pacific group has a lower genetic diversity and a higher rate of inbreeding than the Atlantic group. These results suggest that the Pacific coconuts of Colombia belong to the pre-Columbian population found on the Pacific coast of Panama and Peru. If it had been introduced after Columbus (as in Mexico), genetic diversity would have been higher than on the Atlantic coast.