基于主成分分析和聚类分析的鲜食大豆审定品种综合评价研究  

Comprehensive Evaluation of Approved Fresh Soybean Cultivars Based on Principal Component Analysis and Cluster Analysis

作　　者：李清华[1] 顾智炜 林海峰[1] 颜墩炜 郑龙 陈子琳 柯庆明[1] Qinghua;GU Zhiwei;LIN Haifeng;YAN Dunwei;ZHENG Long;CHEN Zilin;KE Qingming(Putian Institute of Agriculture Science in Fujian,Putian,Fujian 351106;Putian Agricultural Technology Extension Station in Fujian,Putian,Fujian 351100)

机构地区：[1]莆田市农业科学研究所,福建莆田351106 [2]莆田市农业技术推广站,福建莆田351100

出　　处：《中国农学通报》2024年第30期17-25,共9页Chinese Agricultural Science Bulletin

基　　金：福建省科技计划项目-农业引导性(重点)项目“鲜食大豆种质资源收集鉴定与新品种选育”(2021N0039);莆田市科技计划项目—技术开发与应用“优质高产鲜食大豆品种筛选与种质创新”(2023NJJ002)。

摘　　要：为提升福建省鲜食大豆的生产效益与品种优化,本研究对2003—2022年间审定的22个鲜食大豆品种进行了系统性农艺性状分析。通过测定14个主要农艺性状,包括茎粗、采收日数、有效分枝数等,并运用遗传变异分析、相关性分析及主成分分析方法,揭示了性状间的显著相关性和变异规律。结果表明,14个主要农艺性状间存在着丰富变异,变异系数介于3.36%~24.65%之间。茎粗与采收日数呈极显著正相关;有效分枝数、底荚高度、二粒以上标准荚数与株高呈极显著正相关;有效分枝数、二粒以上标准荚数、平均单荚粒数、单株荚重与茎粗呈极显著正相关;单株有效荚数、二粒以上标准荚数与主茎节数呈极显著正相关;每kg标准荚数与标准荚长、鲜百粒重呈极显著负相关。主成分分析,荚产量因子、标准荚饱满度因子、株型因子、荚型因子、生育期因子、标准荚数因子等前6个主成分因子累计方差贡献率达89.15%,基本反映了22个品种原始数据的信息量,可以此构建综合评价模型筛选综合性状得分较高品种。利用类平均法在欧氏距离4.86处可将22个品种划分为5大类群,主成分综合得分较高品种大多处于第Ⅲ类群。通过综合分析评价,本研究筛选出了‘闽豆9号’、‘兴化豆5号’、‘闽豆10号’、‘交大11’、‘闽豆7号’、‘兴化豆1号’及‘兴化豆618’等7个品种供福建省鲜食大豆主要生产应用和育种亲本选择。本研究不仅丰富了福建省鲜食大豆品种的资源库,也为未来的品种改良和产业发展提供了科学指导。The aims are to make better use of the approved varieties of Fujian fresh soybean,and select more excellent varieties for the production and application of Fujian Fresh Soybean and the select breeding parents.In this study,14 main agronomic traits of 22 approved fresh soybean varieties in Fujian Province from 2003 to 2022 were investigated and determined using genetic variation analysis,correlation analysis and principal component analysis,and principal component clustering and comprehensive evaluation were carried out.The results showed that there were abundant variation among the 14 main agronomic traits,and the coefficient of variation ranged from 3.36%to 24.65%.There were significant positive correlations between stem diameter and harvest days,between effective branch number,pod height,standard pod number of two or more seeds and plant height,between effective branches,more than two standard pods,average number of seeds per pod,pod weight per plant and stem diameter,between effective pods per plant,more than two standard pods and node number of main stem.There were significant negative correlations between the standard pod number per kg and the standard pod length,fresh 100-seed weight.Principal component analysis indicated that the cumulative variance contribution rate of the first six principal component factors,such as pod yield factor,standard pod plumpness factor,plant type factor,pod type factor,growth period factor and standard pod number factor,was 89.15%,which basically reflected the information of the original data of 22 varieties,so a comprehensive evaluation model could be constructed to screen the varieties with higher comprehensive character scores.The 22 varieties could be divided into 5 groups at the Euclidean distance of 4.86 by using class-average method,and the varieties with higher comprehensive scores of principal components were mostly in the third group.Through comprehensive analysis and evaluation,seven varieties of‘Mindou 9’,‘Xinghua dou 5’,‘Mindou 10’,‘Jiaoda11’,‘Mi

关 键 词：主成分分析 聚类分析 鲜食大豆 审定品种 品种筛选 品种改良 综合评价 农艺性状 遗传变异 

分 类 号：S323[农业科学—作物遗传育种]