Integrated ML-Based Strategy Identifies Drug Repurposing for Idiopathic Pulmonary Fibrosis.

Area of Science:

• Biomedical research
• Computational drug discovery
• Pharmacology

Background:

• Idiopathic pulmonary fibrosis (IPF) affects 3 million globally, causing lung scarring and progressive respiratory disease.
• Limited FDA-approved treatments and high costs of de novo drug development necessitate alternative strategies like drug repurposing.
• Current drug repurposing methods using computational tools yield low hit rates.

Purpose of the Study:

• To develop an integrated machine learning strategy for efficient and accurate drug repurposing for IPF.
• To identify potentially repurposable drugs and drug combinations for IPF treatment.
• To validate identified drugs through literature, Gene Set Enrichment Analysis (GSEA), and pathway analysis.

Main Methods:

• An integrated machine learning strategy combining predock predictions, literature review, GSEA, and pathway analysis was developed.
• The strategy screened 1480 FDA-approved drugs and drugs in clinical trials against key IPF-related proteins.
• Validation involved assessing existing IPF/fibrosis drug use and predicting involvement in IPF pathways.

Main Results:

• The strategy identified 247 potential drugs, with 27 deemed most promising for IPF repurposing.
• Validation confirmed 72 drugs have been tried for IPF, 13 used for lung fibrosis, and 20 for other fibrotic conditions.
• Pathway analysis implicated 29 pathways in IPF, with 11 directly involved, and suggested 15 synergistic drug combinations.

Conclusions:

• The developed machine learning strategy significantly enhances the accuracy of drug repurposing for IPF.
• The study identified promising drug candidates and combinations for accelerating IPF therapeutic development.
• This approach offers a feasible and efficient method for discovering treatments for IPF and related fibrotic diseases.