Integrated stability analysis of compact cotton genotypes for yield and plant architecture under rainfed conditions using AMMI, GGE biplot, WAAS, BLUP, and MTSI

Background The development of high-yielding stable cotton genotypes with compact plant architecture is critical for improving productivity and sustainability in rainfed ecosystems of India. This study evaluated 22 compact to semi-compact genotypes, including two check varieties (ARBC1651, CSH3075), across three distinct environments in Tamil Nadu during Kharif 2024. Results Genotypic performance was analyzed using a suite of stability assessment tools including AMMI, GGE biplot, WAAS, BLUP, and MTSI. Joint ANOVA revealed highly significant (p ≤ 0.001) genotype, environment, and GEI effects for most traits, with genetic variance contributing up to 72% of phenotypic variation in seed cotton yield (SCY). The first two IPCs of AMMI and GGE biplots explained > 75% of GEI variation and consistently identified Suraksha, Subiksha, and Suraj as stable, high-yielding genotypes. Whereas RS2818, CO17, and Nano represented compact types with plant height < 90 cm, reduced internode length (~ 5.7 cm), and sympodial branching (~ 18.8 cm), aligning with ideotypes suitable for high-density planting and mechanized harvesting. BLUP analysis showed high heritability for SCY (h²mg = 0.91), with strong genotypic variance (71.99%) and high selection accuracy (0.95). It confirmed AMMI and GGE biplots results, identifying Suraksha, Subiksha, and Suraj as superior yielders, while CO17, Nano, and TVH002 promising compact types with high predictive accuracy. WAAS/WAASBY indices confirmed Suraksha and Subiksha as the most stable and productive genotypes, whereas MTSI ranked Suraj (MTSI = 2.16), Subiksha, F2383, and RS2818 as top multi-trait stable performers. Compact plant traits were found to be negatively correlated with excessive vegetative growth but positively associated with yield stability under rainfed conditions. Among the test environments, Veppanthattai (E3) effectively discriminated genotypes, while Coimbatore (E1) and Srivilliputhur (E2) were representative for selecting compact and stable high-yielding types. Conclusion This integrated stability models enabled precise identification of compact, high-yielding genotypes, supporting breeding of rainfed-adapted and machine-harvestable genotypes.