Integrative multi-omics analyses reveal species-specific microbiome–metabolite patterns across the leaves, tubers, and rhizosphere of five yam (Dioscorea) species

Plant–metabolite–microbe interactions are crucial for crop health and quality. Yam ( Dioscorea spp.) is an economically significant crop with both medicinal and edible value. However, the mechanisms by which different yam species regulate metabolite–microbiome interactions across plant compartments remain unclear. In this study, we systematically investigated microbial communities, metabolic profiles, and their potential associations in the leaves, tubers, and rhizospheres of 47 cultivars representing five commonly cultivated Dioscorea species in China. Bacterial communities were mainly shaped by deterministic processes, whereas fungal communities were governed by stochastic processes. Endophyte diversity in leaves and tubers was lower than that in the rhizosphere. We identified 33, 15, and 50 core amplicon sequence variants (ASVs) and 68, 41, and 51 biomarkers from the leaf, tuber, and rhizosphere, respectively, with bacterial biomarkers outnumbering fungal ones. LC-MS analysis identified 1526 metabolites in leaves and 580 in tubers, and revealed clear species-specific patterns. A random forest model accurately discriminated species (out-of-bag error = 0.0426) based on metabolite composition. Co-occurrence network analysis further indicated that specific core microorganisms or biomarkers were significantly and positively correlated with species-specific medicinal metabolites. This study is the first to characterize interactions among closely related Dioscorea species, their associated microbiomes, and metabolites. Our findings provide new insights into the assembly of plant-associated microbiomes, highlight key microbial taxa that may regulate the biosynthesis of functional metabolites, and offer potential targets for improving the health and quality of medicinal and edible yam crops. • First study of metabolite–microbe interactions in 47 cultivars of 5 yam species. • Coexistence of core and species-specific microbe across yam species. • Bacterial and fungal communities are driven by different assemble processes. • Accurate discrimination of yam species by a random forest model based on metabolites. • First identify core and biomarker microbes linked to yam medicinal metabolites.