Study On Methods Of Constructing Core Collection Of Germplasm And Their Applications In Core Construction

Core collection is a representative subset of entire germplasm collection, consisted of limited accessions with minimum genetic redundantcy and retained most genetic diversity in initial collection. The definition and study on core collection is mainly to promote management and utilization of genetic diversity. Accurate measure in genetic similarity among accessions and specifying an appropriate sampling stratey is a key step in developing a core collection. This dissertation will focus on evaluating the properties of different sampling strategies by comparing the genetic variation captured by subsets of upland cotton (Gossypium hirsutum L.), setting up a core collection of island cotton (Gossypium barbadense L.) based on five fiber traits, developing statistical methods of integrating molecular markers and quantitative traits for constructing cores. The main contents and conclusions of this study were as follows:1. A genetic model including environmental effect, row effect, column effect, genotypic effect, genotype by environment effect was employed for predicting genotypic values on five fiber traits of upland cotton (2.5% length, uniformity, strength, elogation, and Micronaire). Mahalanobis distance among accessions based on predicted genotype values were used in classification of the entire collection by UPGMA (unweighted pair group method with arithmetic average) of hierarchical clustering. After specifying the appropriate threshold value of classification with help of the dendrogram, all genotypes could be ranked into some different groups. The standard deviation for each individual was calculated. Forty eight core accessions, with larger standard deviations, were sampled from each group. In terms of the mean, variance, range, and coefficient of variation of traits, the genetic variation captured by the core was evaluated. The results showed that the core could represent the initial collection in genetic diversity; the deviation sampling stratey is an effective method in practice of developing a core collection.2. According to the procedure of stepwise clustering proposed by Hu et al. (2000) for sampling a core collection, a serials of subsets were sampled at 30% proportion, respectively, by different combining scheme of two genetic distances (Mahalanobis distance, Euclidean distance), seven hierarchical cluster methods (Single linkage, Complete linkage, Median method, Centroid method, UPGMA, Weighted pair-group average (WPGMA), Ward's method) and three sampling methods (random sampling, preferred sampling, deviation sampling). The genetic variation of quantitative traits among subsets were compared by evaluating the means,variances, ranges and coefficients of variation of traits. The results showed that the Mahalanobis distance was much better than Euclidean distance in constructing core collection; the preferred sampling and deviation sampling had similar efficiency in increasing the variance and coefficient of variation of subsets when the single linkage was applied in clustering, the preferred sampling is helpful to maintain genotypes with extreame genotypic values; among the seven hierarchical cluster methods, single linkage was the best linkage rules for constructing core collection, which could capture the most genetic diversity of quantitative traits, the succeeding methods were the median method, centroid method and UPGMA.3. A genetic model, including effects of environments, genotypes, and genotype by environment interaction, was employed to analyze five fiber traits of Island cotton (Gossypium barbadense L.). Genotypic values of 304 accessions were predicted by the adjusted unbiased prediction (AUP). Genetic similarities between different accessions were measured by Mahalanobis distances based on genotypic values. Appropriate sampling strategies, linkage rules in stepwise clustering, and sampling proportion were evaluated. To form a core collection of Island cotton, 60 accessions were sampled by the deviation sampling strategy combined with single linkage rule of hierarchical clustering. The genetic variation and structure captured by the core collection were examined in means, variances, ranges and coefficients of variation, correlation coefficients of quantitative traits, and the accessions distribution plotted by first two principal components between two collections. It was showed that the initial collection was well represented by the core collection for exploiting the Island cotton germplasm.4. Two statistical strategies of integrating molecular marker and quantitative traits were proposed for constructing core collection. Material of 2321 soybean accessions was analyzed as a case study for evaluating the strategies. Under 7 sampling proportions, three kinds of subsets were sampled by the stepwise clustering combined with random sampling, based on molecular information, quantitative traits, and integrated information, respectively. The molecular marker diversity was compared in terms of ratios of diversity indice (percentage of polymorphic loci, average expected heterozygosity, average genetic diversity index, average effective number of alleles), and percentage of allelic frequency differences; while, genetic diversity of quantitative traits by mean difference percentage, variance difference percentage, coincidence rate of range, variable rate in coefficient of variation, and variation of single quantitative trait by the ratios of mean, variance, range, and coefficient of variation between subsets and the entire collection. The results showed that the subsets based on molecular markers could increase molecular diversity but genetic variation of quantitative traits; on the contrary, subsets from quantitative...