Genetic diversity and multi-trait selection using the multi-trait genotype-ideotype distance index in a proso millet (Panicum miliaceum L.) germplasm collection

Abstract

Proso millet is a climate-resilient, nutrient-dense nutri-cereal the genetic improvement of which has been constrained by single-trait selection and the limited utilisation of available germplasm. The present study assessed the genetic variability, trait associations and genetic diversity of 259 proso millet genotypes and employed the multi-trait genotype-ideotype distance index (MGIDI) to identify superior genotypes for grain yield and yield-related traits. The genotypes were evaluated over two seasons (2024 and 2025) under well-watered condition, using restricted maximum likelihood (REML), Pearson correlation, Ward's clustering and MGIDI. Highly significant genotypic variation was recorded for all five traits. Broad-sense heritability was high for grain yield (0.834) and thousand-seed weight (0.805); grain yield also recorded the highest genotypic coefficient of variation (48.27 %) and the largest genetic advance as percentage of mean (90.93 %) and this combination indicates high potential for direct selection. Grain yield was positively and significantly correlated with panicle length (r = 0.636) and productive tillers per plant (r = 0.490). The genotypes formed four divergent clusters, with grain yield being the most discriminating trait (η2 = 60.3 %). The MGIDI retained two factors explaining 65.9 % of the variation at a 15 % selection intensity, 39 superior genotypes were identified, resulting in positive predicted gains for all five traits, including a 65.2 % increase in grain yield (6.14 to 10.15 g plant-1). The selected elite genotypes coincided with the high-yielding cluster, reinforcing their reliability as parents for proso millet improvement.