Selecting stable rice mutants with linear mixed models (LMM) and stability indexes

Mutation serves as a pivotal source of diversity in plant breeding. This study focused on identifying stable rice mutant lines. Fourteen rice mutant lines, along with four conventional cultivars, were evaluated in a randomized complete block design with three replicates across three Iranian locations (Rasht, ChaparSar, and Fars province) during two growing seasons (2015, 2016). All statistical analyses were performed using the 'metan' (multi-environment trial analysis) R package. Single-environment ANOVA indicated significant genotypic effects for all traits. Likelihood ratio tests (LRTs) confirmed significant environment and genotype-by-environment interaction (GEI) effects for all traits. The first three principal components (PCs) captured 68.13, 14.46, and 9.76 % of the GEI variation, respectively. Heatmap visualization of yield performance and WAASB (weighted average of absolute scores based on best linear unbiased prediction, BLUP) highlighted genotypes G3, G9, G6, G12, and G5 as both high-yielding and stable. Multi-trait stability index (MTSI) analysis, designed to reveal genotypic strengths and weaknesses, selected only genotypes G7, G5, and G1. The top five genotypes based on the harmonic mean of the relative performance of genotypic values (HMRPGV) were G5, G12, G7, G2, and G1. The superior performance of certain mutants demonstrates that mutation has effectively generated significant genetic diversity. Notably, genotypes G12, G5, and G9 exhibited a clear advantage over the other genotypes and warrant consideration for selection or cultivar release; however, only G5 was selected based on all traits in the MTSI index and could therefore undergo selection or cultivar introduction processes.