Metabolomics-based prediction model for diabetes: A comprehensive analysis of biomarkers and machine learning approaches
Authors' conclusion
Metabolomics data can effectively predict diabetes status, with logistic regression providing the optimal balance of performance and interpretability. The identified metabolites offer potential biomarkers for early diabetes detection and monitoring. This model could serve as a valuable tool for clinical risk assessment and personalized preventive interventions.
Study context
- Population
- Qatari nationals in the Qatar Biobank with metabolomics profiles, instrument measurements and clinical diagnoses42 with diabetes
- Participants
- 450 participants
- Exposure
- Five machine learning models for diabetes prediction from metabolomics data
- Outcome
- Predictive performance for diabetes status (accuracy, precision, recall, F1 score, ROC AUC)
Methodological detailsmachine learning prediction model, Qatar Biobank
- Machine learning prediction model
- Qatar Biobank
Main finding
As reported in the abstract
Prediction of diabetes status
| Group | Accuracy | F1 score | ROC AUC |
|---|---|---|---|
| Logistic Regression | 93.3 % | 0.625 | 0.941 |
Logistic Regression
Accuracy93.3 %
F1 score0.625
ROC AUC0.941
Observational data: associations, not effects.
Further findings
Diabetes prevalence
- Diagnosed with diabetes
- 9.33 % (n = 42)
Metabolites associated with diabetes status
- Metabolites significantly associated (p < 0.05)
- 140
- Glucose
- r = 0.281, p < 0.0001
- Mannose
- r = 0.247, p < 0.0001
- 1,5-anhydroglucitol
- r = −0.297, p < 0.0001
Original publication
Metabolomics-based prediction model for diabetes: A comprehensive analysis of biomarkers and machine learning approaches. Diabetes Res Clin Pract. Published 5 August 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →