Plasma proteomic signatures for type 2 diabetes and related traits in the UK Biobank cohort
Diabetes Research and Clinical Practice, 2025
Gupte T., Azizi Z., Kho P., Zhou J., Nzenkue K., Chen M., Panyard D., Guarischi-Sousa R., Hilliard A., Sharma D., Watson K., Abbasi F., Tsao P., Clarke S., Assimes T.
| Disease area | Application area | Sample type | Products |
|---|---|---|---|
Metabolic Diseases | Pathophysiology Patient Stratification | Plasma |  Olink Explore 3072/384 |
Abstract
Objective
The plasma proteome holds promise as a diagnostic and prognostic tool that can accurately reflect complex human traits and disease processes. We assessed the ability of proteins to predict type 2 diabetes and related traits.
Study design
We analyzed clinical, genetic, and proteomic data from three UK Biobank subcohorts for associations with truncal fat, estimated maximum oxygen consumption (VO2max), and type 2 diabetes. Using least absolute shrinkage and selection operator (LASSO) regression, we compared predictive performance of each trait between data types. The benefit of proteomic signatures (PSs) over the type 2 diabetes clinical risk score, QDiabetes was evaluated. Two-sample Mendelian randomization (MR) identified potentially causal proteins for each trait. Results: LASSO-derived PSs improved prediction of truncal fat and VO2max over clinical and genetic factors. We observed a modest improvement in type 2 diabetes prediction over the QDiabetes score when combining a PS derived for type 2 diabetes that was further augmented with fat- and fitness-associated PSs. Two-sample MR identified a few proteins as potentially causal for each trait.
Conclusion
Plasma PSs modestly improve type 2 diabetes prediction beyond clinical and genetic factors. Candidate causally associated proteins deserve further study as potential novel therapeutic targets for type 2 diabetes.