GDF15 Biomarker and Complication Predictors from UK Biobank Proteomics

50,021 UK Biobank participants without type 2 diabetes were tracked for an average of 12 years to identify protein and metabolite predictors of complications. The research highlights plasma protein GDF15 as a strong predictor for kidney complications, outperforming traditional markers like blood glucose and HbA1c. Ming Hao published this analysis under a CC-BY-4.0 license in May 2026.

Use Cases

• Build predictive models for diabetic kidney complications based on GDF15 protein levels.
• Compare the predictive power of novel omics markers against traditional clinical indicators like HbA1c.
• Apply machine learning methods like LASSO Cox and LightGBM to identify new biomarkers for microvascular and macrovascular complications.
• Use SHAP methods to explain feature contributions within complex predictive models for clinical outcomes.

Strengths

• Includes data from over 50,000 participants from the UK Biobank.
• Analyzes more than 1,400 proteins and 280 metabolites per participant.
• Features a long-term follow-up period averaging 12.0 to 12.03 years.
• Evaluates nine specific machine learning methods for model building.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying dataset is unknown, which may limit suitability assessment.
• The primary file format is a 7.5 MB PDF, which may not contain the raw, structured data.

Provenance

Source

UK Biobank

Collection Method

Prospective study analyzing proteomic and metabolomic data.

Time Range

Baseline from 2006 to 2010, with follow-up averaging 12.0 to 12.03 years.

Freshness

Last updated 2026-05-11 04:26:07; freshness should be verified.

Geography

United Kingdom