Untargeted metabolomics for candidate marker nomination during late-stage cocoa fermentation: a framework toward...

The cutting test remains the conventional standard for determining the endpoint of cocoa fermentation. However, it is subjective, operator-dependent, and lacks biochemical precision. To explore molecular criteria that can complement this traditional approach, we applied an untargeted, multi-platform metabolomics (GC-MS and LC-MS) to characterize biochemical transitions during the last 48 h of fine-flavor cocoa fermentation. Metabolic fingerprints suggest a stage-dependent separation driven by polar and moderately polar metabolites, while lipid profiles remained largely unchanged. Twenty-nine confidently annotated compounds showed consistent temporal dynamics across all platforms, representing sugars, organic acids, peptides, polyphenols, and nitrogenous bases. Using rigorous criteria, including fold change behavior, statistical significance, annotation confidence, and reproducibility, we selected twelve metabolites as candidate indicators whose trends robustly distinguish the late stages of fermentation. This study presents a reproducible workflow for nominating candidate markers of cocoa fermentation completion, which can be transferred, expanded, and validated across different fermentation systems.