Genetic diversity in a citrus germplasm collection characterized with simple sequence repeat markers

The Citrus Variety Collection (CVC), located at the University of California, Riverside contains germplasm accessions within the genus Citrus and in 27 of the 32 related genera in the subfamily Aurantiodeae of the plant family Rutaceae. Genetic diversity and phylogenetic relationships of 370 mostly sexually derived citrus accessions obtained from the CVC were assessed from data on 25 simple sequence repeat markers (SSR). A total of 296 alleles were detected with an average of 11.84 alleles per locus and mean polymorphic information content of 0.633. Phylogenetic relationships among the Citrus accessions were determined by constructing a neighbor-joining tree. Additionally, individual Citrus accessions were probabilistically assigned to populations or multiple populations if their genotype indicated admixture using a model-based clustering approach. Both analyses support the hypothesis that there are only a few naturally occurring species of Citrus and many other Citrus accessions are hybrids of these naturally occurring forms.; Two models are commonly used for the population genetic analysis of simple sequence repeats: stepwise mutation and an infinite allele model. Sixty-five microsatellite alleles from three SSR loci were cloned and sequenced to determine their mode of evolution. Changes in allele size at each SSR marker were due to changes in repeat length in all but one case, a pattern consistent with the stepwise mutation model. The sequence data was utilized to determine relationships between alleles of different taxa. Some point mutations were detected in the flanking regions of the microsatellite that were informative about the evolutionary relationships among alleles. Size homoplasy was apparent at one locus suggesting that phylogenetic inferences should be interpreted cautiously.; A core collection, which is a small subset containing the majority of the genetic diversity, was constructed utilizing SSR marker data. Several different sampling strategies were compared in the construction of the core subset. In relation to the CVC, these different sampling strategies retained 82.4% to 85.8% of the alleles found in the CVC. The subset identified from a proportional stratified sampling strategy retained the largest number of alleles and was most similar in allele frequency to the studied portion of the whole collection.