Statistical and genetic properties of core collection of rice (Oryza sativa L.) derived using advanced strategy and principal component analysis-based methods
2008
Balba, M.L.D.
Three methods of generating core collections were applied to 531 Philippine accessions of Oryza sativa L. The first method used a new software, PowerCore, developed by Kim et al. (2007) which implements the advanced M strategy using a modified heuristic algorithm to establish a core by representing all but minimizing the redundancy of alleles. The other two methods namely, Principal Component Scoring (PCS) and Cluster Analysis with Principal Component Scoring (CA-PCS) are based on multivariate statistical methods: Cluster Analysis and Principal Component Analysis. PCS allows inclusion of entries into the core based on their relative contribution to the over-all diversity, while CA-PCS groups, the accessions into clusters before applying the PCS to each cluster. The core sets derived from each cluster are then combined to comprise the core. To eliminate accessions with similar characteristics, the Gower similarity coefficient was used. Pairs of accessions with a coefficient of at least 0.8 were considered as duplicate and one of each pair was dropped from the core. The final core sets selected by the Powercore, PCS and CA-PCS consist of 46, 55 and 77 accessions; respectively. No redundant accessions were detected in the final core sets. The PowerCore covers all alleles present in the base collections and has the highest average Shannon-Weaver and Nei diversity indices. The three core collections have Mean Difference Percentage less than 20% and Coincidence Rate greater than 80%, hence captured a very high pattern of variation in the base collection
显示更多 [+] 显示较少 [-]