Development of gene-based simple sequence repeat markers for association analysis in Cocos nucifera

Simple sequence repeat (SSR) markers have been widely used to evaluate genetic resources and assist crop breeding in plant systems: they are codominant, easy to score and have abundant polymorphisms. In this study, we used coconut transcriptome sequences to determine the distribution characteristics of gene-based microsatellite loci in Cocos nucifera L. and to develop a series of polymorphic SSR markers. A total of 6,608 SSRs were identified across 57,304 unigenes from coconut transcripts: approximately one expressed sequence tag-derived SSR locus per 10 unigenes. We developed 309 primer pairs and characterized 191 polymorphic gene-based SSR markers across a set of ten coconut accessions from different countries. A total of 615 alleles were identified from the 191 polymorphic SSR loci, with an average of 3.22 alleles per locus. These gene-based SSR markers displayed moderate levels of polymorphism, with observed heterozygosity ranging from 0.18 to 0.695 (mean = 0.385 +/- 0.14). Subsequently, 80 polymorphic markers were selected to assay genotypes of 82 coconut accessions collected from different regions in Asia. Association analysis found a total of nine markers significantly associated with coconut height (P < 0.05) using general linear models in a selective unstructured population and mixed linear models in a mixed-structure population. Two SSR markers (from Unigene9331 and Unigene3350) were significantly associated with coconut height using both statistical methods.