Developing a core collection of litchi (Litchi chinensis Sonn.) based on EST-SSR genotype data and agronomic traits

Abstract Core collections are useful tools to characterize, explore, and exploit the genetic diversity of large germplasms. In this study, we analyzed a litchi germplasm of 96 accessions using the PowerCore program to construct two core collections: core1 containing 18 entries based only on genotype data from 30 EST-SSRs, and core2 of 22 entries based on a combined dataset of the EST-SSRs and 16 phenotypic traits. The core2 collection outperformed the core1 collection by most metrics. All 282 alleles for the 30 markers in the initial germplasm were captured in the core2 collection, which had 100% coverage of alleles. The core2 collection had average Nei diversity and Shannon–Weaver diversity indices of 0.330 and 0.501, respectively, which were slightly higher than those of the initial collection (0.311 and 0.477). Four statistical indices were used to evaluate how well the core collection represented the germplasm: the core2 collection had 0% mean difference, 63.64% variance difference, 128.80% variable rate, and 94.62% coincidence rate of the initial collection. The core2 collection had 100% coverage of the variation in the 16 phenotypic traits, indicating excellent representation of the initial collection. We suggest that when markers have low coverage of the genome, core collections should be based on both genotypic and phenotypic data, which together better represent the variability of the germplasm than genotypic data alone.