Abstract / Summary
Metabolic diseases such as type 2 diabetes (T2D) arise through complex interactions between physiological, molecular, and environmental processes. However, molecular studies often examine traits such as age, sex, adiposity, and glycaemic status independently, overlooking how their combined effects shape molecular phenotypes and relationships across molecular layers. Here, integrating transcriptomic, proteomic, metabolomic, and genetic data from 3,027 individuals in the IMI DIRECT cohort, we investigated the joint effects of age, sex, BMI, and HbA1c on molecular phenotypes and cross-omic relationships. We identified widespread molecular associations alongside extensive context-dependent effects, demonstrating that the effect of one trait frequently depended on the state of another, with sex-dependent age effects particularly prominent. These interactions were also evident in cross-omic relationships, indicating that physiological and metabolic context shapes not only molecular abundance but also the coordination between transcripts, proteins, and metabolites. Probabilistic causal inference further identified directionally structured relationships across molecular layers, with age-associated effects more frequently supporting dependent models consistent with inter-omic mediation. Together, our findings demonstrate that metabolic disease-relevant traits jointly shape molecular phenotypes and cross-omic network organisation. This context dependence highlights the importance of accounting for physiological state when interpreting molecular mechanisms and biomarkers, and provides a framework for investigating dynamic molecular signatures that may inform future approaches to disease stratification and precision medicine.