基于主成分分析和聚类分析的花生主要表型性状综合评价

Comprehensive Evaluation of Major Phenotypic Traits in Peanut Based on Principal Component Analysis and Cluster Analysis

为了构建花生种质资源多性状的综合评价方法,为筛选优异花生种质提供依据,以23个花生品种为研究对象,通过相关分析、主成分分析及聚类分析等方法对花生的17个表型性状进行综合评价。结果显示:17个表型性状的平均变异系数为28.9%;生育期、出仁率、粗脂肪含量、粗蛋白含量等4个性状较稳定,主茎高、第一侧枝长、单株饱果数、单株生产力、籽粒产量、荚果产量、油酸含量、亚油酸含量、油亚比(油酸/亚油酸)等9个性状变异丰富。相关分析表明,荚果产量与籽粒产量极显著正相关,与百果质量、百仁质量、单株生产力、出仁率显著正相关;粗脂肪含量与粗蛋白含量极显著负相关。主成分分析将17个性状综合成2个主成分因子,主要反应产量、油酸和亚油酸含量,累积贡献率达82.301%。通过模糊隶属函数将2个主成分的得分归一化处理后,基于2个主成分的贡献率权重构建了综合评价方程,各花生种质的综合得分范围为-2.840~3.675。在欧式距离为20处,可将23个花生品种划分为3个类群,各类群之间有明显差异。筛选出综合得分(F值)≥1.2的资源材料8份。综合排名显示,漯花15号综合得分最高。综合得分与农艺性状的相关性分析表明,生育期、分枝数、结果枝数、单株饱果数、单株生产力可作为花生品种(系)和花生种质表型综合评价的主要指标。

In order to construct a comprehensive evaluation equation for multiple traits of peanut germplasm resources and provide a basis for screening excellent peanut germplasm,23 new peanut varieties were selected as the research object,and 17 phenotypic traits were comprehensively evaluated through correlation analysis,principal component analysis,and cluster analysis.The results showed that the average coefficient of variation of 17 phenotypic traits was 28.9%.The four personality traits of growth period,kernel yield,crude fat content,and crude protein content are relatively stable,while the nine personality traits of main stem height,first lateral branch length,number of full pods per plant,productivity per plant,kernel yield,pod yield,oleic acid content,linoleic acid content,and the ratio of oleic acid to linoleic acid(O/L)have rich variations.Correlation analysis shows that there is a highly significant positive correlation between pod yield and kernel yield,a significant positive correlation with 100-pod weight,100-kernel weight,individual plant productivity,kernel-pod ratio;The crude fat content is highly significantly negatively correlated with crude protein content.Principal component analysis synthesized 17 personality traits into 2 principal component factors,mainly reflecting yield,oleic acid and linoleic acid content,with a cumulative contribution rate of 82.301%.After normalizing the scores of the two principal components through fuzzy membership functions,a comprehensive evaluation index was constructed based on the contribution rate weights of the two principal components,with a comprehensive score range of-2.840 to 3.675 for each peanut germplasm.At a Standard European distance of 20,23 peanut varieties can be divided into 3 groups,with significant differences between each group Eight resource materials with a comprehensive score(F-value)≥1.2 were selected.The comprehensive ranking shows that Luohua 15 has the highest comprehensive score.The correlation analysis between comprehensive scores and agronomic traits shows that the five agronomic traits of growth period,number of branches,number of fruiting branches,number of full pods per plant,and productivity per plant can be used as comprehensive evaluation indicators for peanut varieties(lines)and as the main indicators for comprehensive evaluation of peanut germplasm phenotypes.