Soybean selection in Kenya enhanced by multi-trait and genotype-by-environment interaction modeling

The growing global demand for soybean protein is driving the expansion of cultivation into new agricultural frontiers. Kenya has been progressing in the development of soybean genotypes to identify those best adapted to its diverse agroecological conditions. However, the selection of genotypes with superior agronomic traits and high stability remains limited. This study quantified the magnitude of genotype-by-environment (G×E) interaction and selected genotypes with the best performance and stability based on a multi-trait approach. A total of 65 genotypes were evaluated across 19 environments from 2019 to 2023/24 growing seasons using a randomized complete block design. Stability was assessed using the Weighted Average of Absolute Scores index, considering plant height (PH), number of days to maturity (NDM), and grain yield (GY). Multi-trait selection was performed using the Desired Gain Index, applying a 20% selection intensity, followed by the genetic gain estimation. Using the likelihood ratio test, we identified significant effects of genotype, environment, and G×E interaction. The overall mean values observed in the experiments were 62.30 cm for PH, 120 days for NDM, and 1783.70 kg ha^-1 for GY. In the multi-trait analysis, we selected the ideotypes G42, G26, G46, G35, G53, G41, G12, G54, G39, G52, G02, and G37. This selection resulted in a genetic gain of 6.5% for PH, 1.3% for NDM, and 15.6% for GY. These genotypes exhibited high genetic potential and adaptation to Kenyan conditions.