Establishment of Core Collections of Japonica Rice Landraces(Oryza Sativa L.) in Taihu Lake Valley

Delin Hong

2007-01-01

Abstract:In order to reduce genetic redundance and improve efficiency of research and utilization of Japonica rice landrace in Taihu Lake Valley,core collections were developed based on the performance of 19 traits of 823 accessions.Based on phenotypic values of the 19 traits,Euclidean distance and Mahalanobis distance were calculated and employed to measure the genetic similarities among the rice landraces.Thirty six tested core collections were first designated by using 6 hierarchical clustering methods(the single linkage,complete linkage,median linkage,centroid method,un-weighted pair-group average and Ward′s method),3 sampling methods(random sampling,preferred sampling and deviation sampling) and 2 genetic distances(Euclidean distance and Mahalanobis distance) at 15% sampling proportion.Six testing parameters(average Shano-Waver index percentage (H′,%),mean difference percentage(MD,%),variance difference percentage(VD,%),coincidence rate(CR,%) for range,variable rate(VR,%) of coefficient of variation and phenotype retained percentage(PR,%),were employed to evaluate the validity of sampling scheme.Further 9 tested core collections were developed using 9 sampling ratio(ranged from 10%-50%) and evaluated for getting appropriate sampling ratio of primary core collection.The results showed that the strategy of using Euclidean distances,un-weighted pair-group average hierarchical clustering methods and preferred sampling was the best sampling scheme and 15%-25% was appropriate sampling ratio of primary core collection.The core collection of Japonica rice landrace in Taihu Lake Valley was established on the basis of Core2,1,5-15(developed by using sampling strategy above) plus 6 lost genotypes of 4 quality traits.This core collection was composed of 129 accessions,accounting for 15.7% of initial collections.Compared with initial collection,the mean of traits showed no significant difference(MD=0),maximum and minimum values of traits coincided,and the genotypes of 4 quality characters were completely remained in the core collection established.