Uncovering rain-fed resilience power of grass pea in Iran using AMMI, BLUP, and multi-trait stability parameters

Rain-fed regions have a low quantity of rainfall with an asymmetric distribution. Therefore, by promoting plants like Lathyrus sativus L., as a legume adapted to unfavorable environments, genotypes with high fodder capacity under such conditions would assist food security worldwide. Here, 16 grass pea genotypes were examined in four rain-fed regions during 2016-2017, 2017-2018, and 2018-2019. Dry fodder yield (DY), plant height (PH), days to flowering (DF), and wet fodder yield (WY) were recorded across 12 test environments. Regarding MLM analysis of variance, LRT_ENV and LRT_ENV×GEN were significant for all studied traits. Phenotypic variance ranged between 1.42 (DY) to 86.9 (PH). Results showed the possibility of grass pea improvement through selection regarding calculated accuracy of selection (> 0.5). PLS regression emphasized the significant role of rainfall during December, January, February, March and April on DY and WY of grass pea. The DY of 16 genotypes across environments varied between 3.4 t/ha (G12 and G16) to 4.6 t/ha (G11). The WY also varied between 16.9 t/ha (G12) and 22.0 t/ha (G8). AMMI analysis revealed G2, and G6 and BLUP-based indices showed G8, and G11 as climate-resilient genotypes with stable DY and WY in rain-fed regions. In this study, WAASB×DY and WAASB×WY plots with equal weights of 50/50 for stability and performance showed G2, G6 as stable genotypes with high DY and WY values. Simultaneous selection based on overall recorded traits using MTSI index addressed G9 > G2 as promising genotypes. Although the polygon view of genotype by yield*trait depicted G1 and G11 as promising grass pea genotypes but G2, and G9 also had positive intermediate superiority indexes without any weakness considering studied traits. It is concluded WAASB×yield > AMMI > BLUP in terms of comprehensiveness in yield stability analysis of grass pea. Also, superiority index as complementary statistics could be incorporated into simultaneous multi-trait stability approaches for achieving exact selection. The identified grass pea genotypes have promising potential in rain-fed regions and could be good candidates for commercial production.