利用基因型值构建宁夏粳稻核心种质的方法

以250份宁夏粳稻为材料,根据7个农艺性状的遗传多样性,研究粳稻核心种质的构建方法.采用混合线性模型统计分析方法无偏预测材料各性状的基因型值,根据2种遗传距离(马氏距离和欧氏距离),3种取样方法(多次聚类随机取样法、多次聚类优先取样法、多次聚类偏离度取样法),8种聚类方法(最短距离法、最长距离法、中间距离法、重心法、类平均法、可变类平均法、可变法、离差平方和法),用均值、方差、极差、变异系数及秩次评价不同遗传距离、取样方法、聚类方法的优劣.结果表明,马氏距离优于欧氏距离；偏离度取样法优于优先取样法和随机取样法,聚类方法最好的是最短距离法、可变类平均法,适宜取样比例是15％.用马氏距离、偏离度取样法、15％取样比例、最短距离聚类法构建了37份宁夏粳稻核心种质.

The Establishment of Ningxia Japonica Rice Core Collection by the Value of Genotype

According to the 7QTL of genetic diversity,used 250 Japonica varieties of Ningxia as materials to study the establishment of Japonica core collection.By statistical analysis technique of best linear unbiased prediction(BLUP),the genotype value of the materials was predicated.The different genetic distance,sampling method and clustering method were estimated by Means,Variance,Range,Coefficient of variation and Rand Analysis Method according 2 genetic distances(Mahalanobis Distance and Euclidian distance),3 sampling method(multi-times random sampling,multi-time preferred sampling and multi-times deviation smapling),8 clustering method(shortest distance method,complete linkage,group average method,barycenter method,UPGMA,varied UPGMA,flexible method,sum of square of the deviations method).The results showed that Mahalanobis Distance was prefered to Euclidian distance,multi-times deviation smapling was prefered to random sampling preferred sampling,the best clustering were the shortest distance method and flexible method,the best sampling ratio was 15%.37 Ningxia Japonica rice core collection were established by Mahalanobis Distance,multi-times deviation smapling,15% sampling ratio and the shortest distance method.