Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DeepDR: an integrated deep-learning model web server for drug repositioning

Shuting Jin, Yi Jiang, Yimin Liu, Tengfei Ma, Dongsheng Cao, Leyi Wei, Xiangrong Liu, Xiangxiang Zeng

TL;DR

DeepDR presents an integrated, open web platform for drug repositioning that unifies six deep learning models across disease- and target-centric tasks. It builds on the DRKG knowledge graph (5.9 million edges, 107 relation types) and a large PubMed corpus to enable rapid, automated predictions without coding. The portal offers input, prediction, and result-visualization components, with interpretable outputs via knowledge-graph paths and drug details to accelerate experimental and computational research. By providing open data, tutorials, and ongoing updates, DeepDR aims to lower barriers to DL-assisted drug discovery and enable collaborative, up-to-date drug repurposing workflows.

Abstract

Background: Identifying new indications for approved drugs is a complex and time-consuming process that requires extensive knowledge of pharmacology, clinical data, and advanced computational methods. Recently, deep learning (DL) methods have shown their capability for the accurate prediction of drug repositioning. However, implementing DL-based modeling requires in-depth domain knowledge and proficient programming skills. Results: In this application, we introduce DeepDR, the first integrated platform that combines a variety of established DL-based models for disease- and target-specific drug repositioning tasks. DeepDR leverages invaluable experience to recommend candidate drugs, which covers more than 15 networks and a comprehensive knowledge graph that includes 5.9 million edges across 107 types of relationships connecting drugs, diseases, proteins/genes, pathways, and expression from six existing databases and a large scientific corpus of 24 million PubMed publications. Additionally, the recommended results include detailed descriptions of the recommended drugs and visualize key patterns with interpretability through a knowledge graph. Conclusion: DeepDR is free and open to all users without the requirement of registration. We believe it can provide an easy-to-use, systematic, highly accurate, and computationally automated platform for both experimental and computational scientists.

DeepDR: an integrated deep-learning model web server for drug repositioning

TL;DR

DeepDR presents an integrated, open web platform for drug repositioning that unifies six deep learning models across disease- and target-centric tasks. It builds on the DRKG knowledge graph (5.9 million edges, 107 relation types) and a large PubMed corpus to enable rapid, automated predictions without coding. The portal offers input, prediction, and result-visualization components, with interpretable outputs via knowledge-graph paths and drug details to accelerate experimental and computational research. By providing open data, tutorials, and ongoing updates, DeepDR aims to lower barriers to DL-assisted drug discovery and enable collaborative, up-to-date drug repurposing workflows.

Abstract

Background: Identifying new indications for approved drugs is a complex and time-consuming process that requires extensive knowledge of pharmacology, clinical data, and advanced computational methods. Recently, deep learning (DL) methods have shown their capability for the accurate prediction of drug repositioning. However, implementing DL-based modeling requires in-depth domain knowledge and proficient programming skills. Results: In this application, we introduce DeepDR, the first integrated platform that combines a variety of established DL-based models for disease- and target-specific drug repositioning tasks. DeepDR leverages invaluable experience to recommend candidate drugs, which covers more than 15 networks and a comprehensive knowledge graph that includes 5.9 million edges across 107 types of relationships connecting drugs, diseases, proteins/genes, pathways, and expression from six existing databases and a large scientific corpus of 24 million PubMed publications. Additionally, the recommended results include detailed descriptions of the recommended drugs and visualize key patterns with interpretability through a knowledge graph. Conclusion: DeepDR is free and open to all users without the requirement of registration. We believe it can provide an easy-to-use, systematic, highly accurate, and computationally automated platform for both experimental and computational scientists.

Paper Structure

This paper contains 20 sections, 16 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. DeepDR platform
  3. The overall framework of DeepDR
  4. DeepDR dataset
  5. DeepDR services and portal
  6. User Inputs module
  7. Prediction module
  8. Result analysis and visualization
  9. Discussion and conclusion
  10. METHODS
  11. Web-server implementation
  12. Overview and Guidance of Disease-Centric Models
  13. DeepDR
  14. HeTDR
  15. DisKGE
  16. ...and 5 more sections

Figures (4)

  • Figure 1: The graph abstract of DeepDR.
  • Figure 2: The overview of the DeepDR platform. (A) DeepDR dataset, (B) DeepDR services, and (C) DeepDR portal screenshot.
  • Figure 3: DeepDR services description. (A) User input module, (B)disease-centric DeepDR service module, (C) target-centric DeepDR service module, and (D)result analysis module.
  • Figure 4: Performance of models built on DeepDR web-server.