Integrated Genomic–Metabolomic Analysis for Tri-Categorical Classification of Type 2 Diabetes Status in the Korean Ansan–Ansung Cohort

Identifying high-risk individuals for type 2 diabetes (T2D), particularly during prediabetes (PD), remains challenging owing to its complex metabolic etiology. In this study, we aimed to develop and validate an integrative multi-omics model for the tri-categorical classification of T2D status (Normal Glucose Tolerance, PD, and T2D) by combining genomic and metabolomic data from a Korean cohort. Based on cross-sectional data from 1819 participants in the Ansan-Ansung cohort, significant metabolites associated with glycemic traits and T2D status were identified using regression analysis. A metabolite-adjusted genome-wide association study (GWAS) was conducted to identify T2D-associated genetic variants. Finally, three nested prediction models (Clinical, Metabolite-Enriched, and integrated Multi-omics) were constructed using baseline-category logistic regression and evaluated using stratified five-fold cross-validation. Thirty-nine metabolites were identified as consistently associated with T2D status and related glycemic traits. GWAS identified 86 T2D-associated independent single-nucleotide polymorphisms (SNPs). The final integrated multi-omics model, combining clinical factors, 39 metabolites, and 86 SNPs, demonstrated strong predictive performance for classifying T2D status, achieving an area under the receiver operating characteristic curve (AUC) of 0.935, significantly improved over the clinical model (AUC = 0.695) and metabolite-enriched model (AUC = 0.874). It also outperformed previously established external models and represents an important step in our understanding of T2D status. Our findings thus demonstrate that integrating genomic and metabolomic data provides a useful framework for the tri-categorical classification of T2D status. This multi-omics approach significantly enhances risk stratification beyond that provided by clinical or single-omics data alone, thus offering valuable insights into the underlying pathophysiology in T2D with potential for shaping future T2D research and clinical practice.