Multi-trait model-based assessment of selection efficiency and stability in soybean germplasm.

Soybean (Glycine max L. Merrill) productivity is strongly influenced by genotype × environment (G × E) interaction, while simultaneous improvement of yield, seed quality, and associated traits remains a major breeding objective. Twenty-two soybean genotypes were evaluated across eight environments representing four locations and two kharif seasons in Maharashtra, India, to identify genotypes combining desirable multi-trait performance, stability, and adaptability. Significant variation was found for all the traits studied, seed yield ranged from 0.26 to 1.68 kg plot- 1 and protein content from 33.25 to 41.83%. Pooled analysis of variance showed highly significant effects of genotype, environment and G × E interaction. Number of pods per plant, plant height and protein content showed positive association with seed yield, while flowering and maturity duration showed negative association with seed yield. Factor analysis of MGIDI explained 71.6% of the total variance with 3 factors. WAASB and WAASBY analyses provided complementary information on genotype stability and performance-stability relationships. The multi-trait selection methods MGIDI, MTSI, MTMPS and FAI-BLUP identified overlapping sets of desirable genotypes and the coincidence analysis showed consistency of the selection criteria. MAUS 725 (G10) and KDS 992 (G13) were consistently identified as promising genotypes, combining favourable seed yield, yield-related traits, seed quality, and stability across the evaluated environments. These genotypes can be used for further multi-environment validation and potential parents in soybean improvement program. The integration framework improved selection by including multi-trait desirability, stability and BLUP based prediction for more reliable breeding decisions.