Identification of Seed Metabolites and Microbiota members associated with Germination and Emergence in Common Bean
Seed quality is a complex trait shaped by morphological, biochemical and microbiological properties that are rarely characterised simultaneously, limiting our ability to identify robust predictive indicators of germination speed and seedling emergence across varieties. Here, we performed a multi-factor characterisation of eight common bean (Phaseolus vulgaris L.) varieties, combining seed morphometrics, untargeted GC-MS metabolomics on three seed organs, and amplicon sequencing of bacterial and fungal communities, to identify indicators of germination speed and emergence percentage. The eight varieties showed substantial variation in both traits, used as physiological seed quality proxies. Seed weight and size variation between varieties were correlated with germination speed. The intravariety variance of seed weight was independently correlated with emergence performance. Metabolome composition differed strongly across seed organs, with variety as the dominant driver. Individual-seed metabolomic profiles in the plumule and cotyledon were associated with germination speed but not emergence, yielding 16 plumule and three cotyledon candidate metabolite markers. Fungal community composition was associated with both germination speed and emergence, while bacterial communities were associated with emergence only. Nine fungal and four bacterial taxa were identified as candidate indicators. Inter-kingdom co-occurrence network analysis revealed that fungi with similar germination speed associations tend to cluster in the same modules, suggesting that community-level modules rather than individual taxa may constitute more robust microbial indicators. These results demonstrate that germination speed and emergence capacity are governed by distinct seed properties, and provide morphological, metabolic and microbial candidate indicators for integration into targeted seed quality assessment frameworks for common bean.