RepurposeDrugs: an interactive web-based database and machine learning predictor for repurposing mono- and combination therapies

RepurposeDrugs (https://repurposedrugs.org/) is a web-based data portal that integrates a comprehensive database and machine learning (ML) based predictor designed to uncover novel drug-disease associations (drug repurposing) for mono and combination therapies. It provides an interactive GUI to explore associations between drugs and diseases, categorized into 21 groups, including neoplasms, skin, cardiovascular and other diseases. There are 575 approved drugs, 55 approved drug combinations, and 219 disease indications in RepurposeDrugs. It provides insights into drug approval status, indications, and information on withdrawn or ongoing clinical trials for mono and combination therapies across various diseases.

In addition to collected data, RepurposeDrugs also provides ML-based predicted drug-disease associations for mono and combination drugs. The backend ML algorithm is trained on known approved/failed indications. It can predict disease association scores for a broad spectrum of compounds uploaded by the end users, regardless of their development status (approved, investigational, or preclinical).

Accessible without any login requirements, RepurposeDrugs offers most features readily available through a mouse click, requiring no programming skills or in-depth biological knowledge. The portal provides data for customized drug-disease associations (known and predicted), which can be exported into a publication-quality figure in PNG or SVG formats and tabular data in XLSX format.

To our knowledge, RepurposeDrugs is the first integrative mono and combination therapies database incorporating an ML-based web tool for predicting the likelihood of treatment approval. Unlike conventional drug repurposing databases, it can covers a broad range of diseases and various single drugs and drug combinations, opening new avenues for researchers and clinicians seeking innovative therapeutic solutions.