甘蔗杂交品种初级核心种质取样策略

Strategy of Sampling for Pre-Core Collection of Sugarcane Hybrid

以国家甘蔗种质资源圃中1202份甘蔗杂交品种为材料,根据23个数量和质量性状,从分组原则、组内取样比例、组内取样方法3个层次探讨构建甘蔗杂交品种初级核心种质的最佳取样策略,共形成26种取样策略;同时设10个总体取样量梯度,确定最佳的总体取样量。分组原则以原产地、种植区域、总体聚类进行分组及不分组的大随机;组内取样比例按组内个体数量的简单比例(P)、平方根比例(S)、对数比例(L)和多样性比例(G)确定;组内取样方法采用聚类(C)和随机(R)2种方法;10个总体取样量梯度为5%、10%、15%、20%、25%、30%、35%、40%、45%和50%。应用变异系数、遗传多样性指数、表型保留比例、表型频率方差、表型方差等5个参数来检验各取样策略的优劣。结果表明,聚类取样优于随机取样;总体聚类分组优于其他分组;在聚类取样中,平方根比例最好,在随机取样中,多样性比例最好;根据取样策略及总体取样量的分析结果最终确认按10%总体取样量,以总体聚类分组、按对数比例在组内聚类取样为构建甘蔗杂交品种初级核心种质的最佳策略组合,其遗传多样性明显高于总资源库。在此初级核心种质的基础上,加入极值材料和取样极易丢失表型性状的材料共计136份组成最终初级核心种质,占总资源的11.31%。

Sugarcane is a major economic crop in the world and sugarcane hybrids provides an important gene pool for sugarcane. The National Nursery of Sugarcane Germplasm Resources （NNSGR） in China has maintained a large collection of sugarcane hybrids from over ten provinces of China and twenty countries around the world. In this study, 1 202 sugarcane hybrids were selected as materials, 26 sampling procedures were used, which consisted of 4 grouping principles （cluster grouping, grouping based on original regions, grouping based on planting regions and not grouping）, 4 sampling proportions within group （simple proportion, logarithmic proportion, square root proportion and genetic proportion）, 2 individual selection methods within group （random selection and cluster selection） on the basis of 23 quantitative and qualitative traits. At the same time, 10 grads for sampling percentage （5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, and 50%） from total resources were analyzed to decide the best one of them. The sampling strategies and the pre-core collection were validated and evaluated with five effective parameters including indices of genetic diversity, variance of phenotypic value, variance of phenotypic frequency, coefficient of variation and ratio of phenotypic traits retained. The result indicated that the cluster sampling was better than the random sampling; cluster grouping was the best grouping principle; square root proportion was the best sampling proportion within group in clustering sampling, genetic proportion was the best in random sampling. Finally, an optimal pre-core collection was set up based on cluster grouping, logarithmic proportion, and cluster sampling according to 10% sampling percentage by analyzing 26 sampling strategies and 10 sampling ratio. The pre-core collection contained 136 accessions, which accounted for 11.31% of total resources and its genetic diversity was better than that of total resources. They included 120 accessions sampled by program and 16 accessions chosen by directional selection, which included the elite ones and some with the rare phenotypic traits.