Principal coordinate scoring procedure in the establishment of a Philippine mungbean (Vigna radiata (L.) Wilczek) core collection

In this study, a procedure for the establishment of a core collection was developed. The developed procedure uses the morphoagronomic characterization of the accessions as criteria in the selection of the core entries. The procedure is divided into three phases, namely: (1) identify the accessions with outlying values on any of the quantitative traits; (2) stratify the germplasm collection based on the quantitative traits; and (3) from each stratum of the germplasm collection, select a subset of accessions that captures the qualitative traits diversity of the stratum. Identification of the outliers facilitates the nomination of the typical accessions. Stratifying the germplasm collection into groups of related accessions and selecting representative subset from each stratum improves the core's captured quatitative traits diversity. On the other hand, the use of the qualitative variables as criteria for the selection of a stratum's representative subset, facilitate the capture of the qualitative trait phenotypes. The principal coordinate scoring (PCOS) procedure was developed to implement Phase 3 of the aforementioned procedure on core establishment. The PCOs procedure uses principal coordinate analysis to come up with the coordinates of the accessions in a reduced Euclidean space. Principal coordinate analysis permits the projection of the accessions on a euclidean space where the original distance relationship, computed from the Gower's similarity coefficient, between the accessions is preserved. Base on the inter-accession distance in the reduced Euclidean space, a subset of accessions that maximizes inter-accession dissimilarity is selected. The procedure on core establishment was used to establish a core from the AVRDC Philippine Mungbean (Vigna radiata) germplasm collection. Evaluation of the established core revealed that with respect to each of the quantitative variables, core entries covered the ranges defined by the minimum, 25th percentile, median 75th percentile and maximum values of germplasm collection. Furthermore, with respect to the qualitative trait phenotypes captured by the core collection, the principal coordinate scoring procedure has been compared with random selection of accessions and was found to be better than the latter