Development and evaluation of a core subset of the USDA rice germplasm collection

A valuable core collection that is a subset of a whole germplasm collection should capture most of the variation present in the whole collection, while allowing for more efficient evaluation and management due to smaller size. The United States Department of Agriculture (USDA) rice (Oryza sativa L.) core subset (RCS), assembled by stratified random sampling, consists of 1790 entries from 114 countries and represents approximately 10% of the 18412 accessions in the rice whole collection (RWC). Data for this study were obtained from the USDA germplasm system at www.ars-grin.gov for the RWC and from an evaluation conducted in 2002 for the RCS. Comparative analysis for frequency distributions of 14 descriptors demonstrated that the RCS was highly correlated with the RWC (r = 0.94, P < 0.0001). Thus, information drawn from the RCS could be effectively used to assess the RWC with 88% certainty. Correlation coefficients between the RCS and the RWC for eight descriptors were >= 0.9, indicating that the RCS was highly representative of the RWC. Correlation coefficients for the other six descriptors were lower (0.65-0.88), but still significant.