Analysis and regulation of baijiu aroma-active compounds: integrating metabolomics, sensory genomics, and artificial intelligence.

Baijiu aroma is governed by a small number of odor-active compounds present at trace or ultra-trace levels, whose formation and perception are shaped by complex microbial metabolism, physicochemical transformations, and sensory interactions. While metabolomics, sensory genomics, and artificial intelligence (AI) have each been applied to Baijiu aroma research, their integration has remained largely fragmented. This review provides a critical and integrative framework that connects chemical profiling, sensory relevance, and data-driven modeling to elucidate aroma-active compound formation and regulation in Baijiu. We synthesize recent advances showing how multi-platform metabolomics enables quantitative mapping of key aroma compounds, how sensory genomics links odor activity values and interaction effects to perceptual outcomes, and how machine learning models facilitate cross-modal data integration for aroma prediction and process optimization. Importantly, this review highlights methodological limitations that currently constrain the field, including insufficient sensitivity for ultra-trace compounds, variability and subjectivity in sensory data, limited generalizability of AI models, and incomplete mechanistic validation of predicted pathways. By explicitly integrating metabolomics, sensory genomics, and AI into a unified analytical framework, this review moves beyond descriptive summaries and provides mechanistic insights and practical perspectives for intelligent aroma regulation, quality consistency, and future industrial application in Baijiu production.