Project Introduction

SgTxGNN: Drug Repurposing Predictions for Singapore


Background

Drug repurposing (also known as drug repositioning) is a strategy for identifying new uses for existing, approved drugs. This approach offers several advantages over traditional drug discovery:

  • Reduced development time: Existing drugs have known safety profiles
  • Lower costs: Much of the clinical work has already been done
  • Higher success rates: Known compounds reduce technical risk

The TxGNN Model

SgTxGNN is built on TxGNN (Therapeutic Target Prediction using Graph Neural Networks), developed by Harvard Medical School's Zitnik Lab and published in Nature Medicine (2023).

Key Features

  • Knowledge Graph: Integrates biomedical knowledge from multiple sources
  • Deep Learning: Uses graph neural networks for prediction
  • Clinical Focus: Designed specifically for drug repurposing

Citation

Huang, K., Chandak, P., Wang, Q. et al. A foundation model for clinician-centered drug repurposing. Nat Med (2023). https://doi.org/10.1038/s41591-023-02233-x


Singapore Adaptation

SgTxGNN adapts TxGNN specifically for Singapore:

Data Sources

Source Description
HSA Health Sciences Authority drug registry
DrugBank Drug-target relationships
TxGNN KG Biomedical knowledge graph

Coverage

  • 745 drugs mapped to DrugBank
  • 31,543 predictions generated
  • 4,589 diseases covered

Methodology

1. Drug Mapping

Singapore HSA drug data is mapped to international identifiers:

HSA Drug Registry → Ingredient Normalisation → DrugBank ID

Mapping rate: 73.87% of HSA ingredients successfully mapped

2. Prediction Methods

Two complementary approaches:

Method Description Count
Knowledge Graph (KG) Direct relationship queries 22,136
Deep Learning (DL) Neural network predictions 29,100
Unified (KG+DL) Dual-validated predictions 1,217

3. Evidence Classification

Predictions are classified by evidence level:

Level Description
L1 Multiple Phase 3 RCTs
L2 Single RCT or Phase 2 trials
L3 Observational studies
L4 Preclinical/mechanistic evidence
L5 Model prediction only

Use Cases

For Researchers

  • Identify promising drug repurposing candidates
  • Prioritise experimental validation studies
  • Explore drug-disease relationships

For Clinicians

  • Research potential off-label uses
  • Understand drug mechanisms
  • Access structured evidence summaries

For Healthcare Institutions

  • Support translational research
  • Enable clinical decision support research
  • Facilitate EHR integration via SMART on FHIR

Limitations

Important considerations:

  1. Research use only: Predictions have not been clinically validated
  2. No clinical guidance: Not intended for treatment decisions
  3. Model uncertainty: AI predictions may be incorrect
  4. Singapore focus: Based on HSA-approved medications

Future Directions

Planned enhancements:

  • Integration of clinical trial evidence
  • Literature-based validation scoring
  • Drug interaction analysis
  • Periodic model updates

Contact


Back to top

Copyright © 2026 Yao.Care. For research purposes only. Not medical advice.

This site uses Just the Docs, a documentation theme for Jekyll.