Multivariate evaluation of purple sweetpotato (Ipomoea batatas L.) genotypes using the MGIDI index

Abstract Purple sweetpotato ( Ipomoea batatas L.) is an important food crop with increasing demand for high-yielding and pest-resistant cultivars to support sustainable production and food security. This study evaluated 135 sweetpotato genotypes derived from true seeds introduced from the International Potato Center (CIP). Each genotype was clonally propagated and field-evaluated to identify lines combining high yield, vigorous growth, and weevil resistance. Six agronomic traits—growth vigor (GV), number of vines harvested (NVH), marketable root number (MRN), marketable root yield (MRY), total yield (TY), and number of weevil-free roots (WFR)—were analyzed using an integrated framework of linear mixed models (LMM), correlation analysis, hierarchical clustering, and the multi-trait genotype–ideotype distance index (MGIDI). Significant phenotypic and genetic variation was detected across genotypes, with high heritability values for traits (0.605–0.965), reflecting substantial genetic variability among genotypes for the traits evaluated. Yield-related traits showed the highest variability and were strongly correlated ( r > 0.9), while WFR was largely independent, suggesting distinct genetic regulation of pest resistance. Cluster analysis classified genotypes into three phenotypic groups, of which Cluster 3 displayed favorable performance for both yield and resistance. Multi-trait genotype–ideotype distance index analysis further identified two major latent factors corresponding to productivity and growth–resistance traits. Integrating the results from all analyses, 15 genotypes were selected as the most promising materials for further validation following this preliminary screening study. Among them, CIP120806.009, CIP120566.517, and CIP120566.452 were identified as the most promising genotypes based on MGIDI ranking and overall multi-trait performance. The findings demonstrate that combining LMM, clustering, and MGIDI provides an effective multi-trait selection approach for identifying desirable sweetpotato genotypes for subsequent validation in sweetpotato improvement programs.