Evaluation and grouping varieties and lines of soybean under normal and drought stress using multivariate statistical methods in two regions of Rasht and Gonbad Kavous

Background and objectives: Considering of drought stress as a major limitation in production and stable yield of crops including soybean, identification of tolerant genotypes to water deficit can be a valuable approach against the problem. Present study was designed to evaluate varieties and advanced lines of soybean based on different agronomic traits to recommendation of superior varieties and lines in normal and under drought stress environments in two regions of Rasht and Gonbad Kavous.
Material and methods: Plant materials including 100 advanced lines with 21 varieties of soybean were planted in two environments under water deficit and normal conditions in two regions in 2014 as randomized complete block design with three replications. Twenty one traits including phonologic, morphologic, yield and yield components were measured. The combined analysis of variance, cluster analysis, discriminant function analysis and principle component analysis were done.
Results: The results of analysis of variance for all of the evaluated traits revealed that difference among genotypes and at least one of the interactions were significant at one percent probability level indicating high genetic diversity and different reaction of genotypes to different conditions for investigated traits. Therefore other analysis including cluster analysis, principle components analysis and discriminant function analysis were performed for each condition separately. The cluster analysis was classified varieties and lines to two or three groups. The superior varieties and lines were in the clusters with means higher than other groups and total mean for important traits. In total varieties and lines 8, 16, 36, 37, 42, 63, 67, 72 and 86 were in the superior group in both normal and drought stress conditions of Rasht region and varieties and lines 34, 43, 62, 66 and 70 were better than others in both normal and drought stress conditions of Gonbad Kavous. Discriminant function analysis based on groups derived from cluster analysis revealed that in the most environments total weight of plant, grain yield per plant, total pod number and seed number per plant were the most effective traits in distinguish of the groups. As the groups including better advanced lines and varieties had higher values foer these traits.
Conclusion: It is expected to use the identified tolerant varieties and lines in the studied regions after verification of the results in serveral years. Considering the role of more efficient some traits including total weight of plant, grain yield per plant and total pod number, those should be prioritized in breeding programs such as selection and improving of genotypes.