N. V. Stepanova, Zh. N. Mukhatova, A. G. Subbotin, N. A. Shyurova
Saratov State University of Genetics, Biotechnology and Engineering named after N. I. Vavilov, Saratov, Russia
E-mail:
Volume 26 No. 5
Date of paper submission: 06.06.2025, date of review: 28.07.2025, date of acceptance: 11.03.2026.
Published: 05/31/2026
Abstract. The purpose id to carry out a breeding analysis of the model soybean population by applying multidimensional statistical methods. Methods. When studying the source material, modern breeding programs take into account a set of economically significant indicators, but when exposed to environmental conditions, the problem of variability of the source material becomes acute, which makes it difficult to evaluate it, and, consequently, to search for genetically valuable forms. In this regard, the processing of research results was carried out using multivariate statistical methods, which helps to identify valuable genotypes more effectively. The study included six varieties from different breeding programs and twenty breeding lines from the Vavilov University. The research results were processed using methods of variance, cluster, and factor analysis. Results. The cluster analysis allowed the study samples to be divided into six cluster groups. The variety sample of the sixth cluster 4-27 No. 160 (2.99 t/ha) with the largest number of beans per plant (19.9 pcs.), the number of seeds and the weight of seeds per plant (29.8 pcs. and 4.5 g), and the fat content in seeds (21.2 %) proved to be the most productive. The constructed correlation matrix showed that seed yield has a positive correlation with the number of beans per plant (0.78), the number of seeds per plant (0.78), the weight of seeds per plant (0.92), and the content of NFE (0.42). Based on the established determination coefficients, the greatest contribution to the total seed yield of soybean varieties was made by such structural elements as seed weight per plant (33.02 %), number of beans per plant (23.73 %), and number of seeds per plant (23.73 %). The use of the principal component analysis method allowed for the identification of eight hypothetical factors that accounted for 95.45 % of the accumulated variance. Scientific novelty. Thus, using multidimensional statistical data processing methods, for the first time, a selection assessment and selection of valuable source material from common and new soybean varieties of various selections and lines created at the Vavilov University for further breeding work with them were carried out.
Keywords: soybeans, variety pattern, multidimensional statistics, cluster analysis, factor analysis, determination
Acknowledgments. The study was completed under the state assignment of the Ministry of Science and Higher Education of the Russian Federation No. 1025042500017-3-4.1.6.
For citation: Stepanova N. V., Mukhatova Zh. N., Subbotin A. G., Shyurova N. A. Analysis of a model soybean population using multidimensional statistical methods. Agrarian Bulletin of the Urals. 2026; 26 (05): 816‒826. https://doi.org/10.32417/1997-4868-2026-26-05-816-826 (In Russ.)
Download the full text of the article