Integrating Multivariate Analysis and DNA Barcoding for Amaranth Germplasm Characterization and Promising Genotype Selection

Amaranth (Amaranthus spp.) is a multifaceted genus of C4 plants with significant nutritional and agronomic potential, yet it remains underutilized in mainstream agriculture. Despite growing interest in Amaranth, most germplasm studies have used either phenotypic or molecular approaches alone, lacking integration. Multivariate methods have not been systematically applied to identify promising genotypes, and species-specific selection indices for grain Amaranth remain unexplored. To address these gaps, this study comprehensively characterized 84 Amaranth genotypes representing multiple species (A. caudatus, A. cruentus, A. hypochondriacus, A. hybridus, A. spinosus, A. powellii, A. tricolor, and 38 accessions of unknown taxonomic status) using field experiments in a randomized complete block design with three replications and DNA barcoding with chloroplast (psbA-trnH) and nuclear (ITS) markers. Analysis of variance revealed highly significant differences (p < 0.01) among genotypes for all six agronomic traits evaluated, confirming substantial genetic variability with grain yield exhibiting the widest variation (CV = 28.55%), ranging from 0.25 to 125.56 g/plant. ... mehrHigh broad-sense heritability estimates (0.79–0.99) coupled with high genetic advance, particularly for grain yield (117.54%), indicated that these traits would respond favorably to selection. Path analysis and stepwise regression identified early flowering, long inflorescences, and heavy seeds as the primary determinants of grain yield, collectively explaining 27% of yield variation. Mahalanobis D$^2$ analysis identified nine multivariate outliers with distinct phenotypic profiles, among which G39 emerged as the most promising breeding candidate, combining exceptional yield (90.50 g/plant) with desirable architecture, long inflorescence, and large seeds. Principal component analysis further resolved trait complexes, identifying 11 PC1-selected promising genotypes as donors for plant architecture and three PC2-selected promising genotypes as donors for seed size characteristics. Molecular analysis revealed distinct genetic relationships. A. caudatus (kiwicha) exhibited limited haplotype diversity indicating a narrow genetic base, while A. cruentus and A. hypochondriacus showed broader diversity, with the nuclear ITS network providing clearer resolution than chloroplast markers due to biparental inheritance. Outlier genotypes, including G82, G83, G13, G10, and G39, occupied unique haplotype positions, confirming that their phenotypic distinctiveness corresponds to genuine genetic differentiation. The novelty of this study lies in integrating multivariate biostatistical techniques (heritability, path analysis, Mahalanobis D$^2$, PCA, and stepwise regression) with two complementary DNA barcode systems (chloroplast and nuclear) within a single germplasm collection. This integrated approach provides breeders with well-characterized germplasm, validated selection criteria, and prioritized parental materials for Amaranth improvement. Further multi-location and multi-season evaluations are recommended to ensure the stability and adaptability of these promising germplasm accessions.