Repurpose Open Data to Discover Therapeutics for COVID-19 Using Deep Learning.

Area of Science:

• Computational biology
• Pharmacology
• Artificial intelligence in medicine

Background:

• The COVID-19 pandemic, caused by SARS-CoV-2, has resulted in millions of cases and deaths globally.
• Effective treatments for COVID-19 are limited, necessitating novel therapeutic strategies.
• Drug repurposing presents a viable approach to accelerate the development of COVID-19 therapies.

Purpose of the Study:

• To develop and apply an integrative, network-based deep-learning methodology for identifying repurposable drugs against COVID-19.
• To create a comprehensive knowledge graph integrating diverse biological and medical data.
• To validate identified drug candidates using transcriptomic, proteomic, and clinical trial data.

Main Methods:

• Construction of a large-scale knowledge graph with 15 million edges from 24 million PubMed publications, encompassing drugs, diseases, proteins, pathways, and gene expression.
• Application of a network-based deep-learning framework on cloud computing resources (AWS) to analyze the knowledge graph.
• Validation of identified drug candidates through transcriptomic and proteomic data from SARS-CoV-2-infected cells and clinical trial information.

Main Results:

• Identification of 41 repurposable drugs with potential therapeutic associations for COVID-19.
• Key identified drugs include dexamethasone, indomethacin, niclosamide, and toremifene.
• Therapeutic associations were supported by experimental data and ongoing clinical trial evidence.

Conclusions:

• The study demonstrates a powerful deep-learning methodology (CoV-KGE) for prioritizing existing drugs for COVID-19 treatment.
• This approach has the potential to significantly accelerate the discovery and development of new therapies.
• While not recommending specific drugs, the methodology offers a valuable tool for future drug repurposing efforts in infectious diseases.