Integrative Proteomic and Machine Learning Analysis Identifies Novel Predictors and Risk Model for Diabetic Macrovascular Complications
Authors' conclusion
This work identifies a robust circulating protein panel suitable for early forecasting of diabetic macrovascular events and sheds light on core biological drivers of disease progression.
Study context
- Population
- Patients with diabetes, for prediction of diabetic macrovascular complications
- Outcome
- Circulating protein biomarkers for diabetic macrovascular complications
Methodological detailsproteome-wide Mendelian randomization, Cox proportional hazards regression, proteome-wide association study, machine learning-derived protein risk score, phenome-wide Mendelian randomization (MR)
- Proteome-wide Mendelian randomization
- Cox proportional hazards regression
- Proteome-wide association study
- Machine learning-derived protein risk score
- Phenome-wide MR
Performance
As reported in the abstract
Proteins identified, clustering in immune-inflammatory and extracellular matrix remodeling pathways
43
12-protein panel discrimination (AUC)
0.793
Protein perturbations emerged before clinical onset
10-12 years
Further findings
Multi-omics integration
- Synchronized links to cardiometabolic dysregulation and cardiac structural alterations
Phenome-wide MR
- Pleiotropic associations across disease categories; causal support for several prioritized proteins
Original publication
Integrative Proteomic and Machine Learning Analysis Identifies Novel Predictors and Risk Model for Diabetic Macrovascular Complications. Diabetes Obes Metab. Published 22 April 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →