Potential Target Discovery and Drug Repurposing for Coronaviruses: Study Involving a Knowledge Graph-Based Approach

Area of Science:

• Biomedical Informatics
• Computational Biology
• Virology

Background:

• Global pandemics caused by coronaviruses (SARS, MERS, COVID-19) highlight the urgent need for effective countermeasures.
• Coronaviruses are continuously evolving, posing an ongoing threat of future outbreaks.
• Knowledge graphs offer a promising avenue for understanding virus pathogenicity and transmission mechanisms.

Purpose of the Study:

• To discover potential therapeutic targets for coronaviruses.
• To identify candidate drugs for repurposing against coronaviruses.
• To leverage a knowledge graph-based approach for accelerated drug discovery.

Main Methods:

• Constructed a comprehensive coronavirus knowledge graph (CovKG) by integrating biomedical literature and existing knowledge bases.
• Developed a semantic conversion model to organize extracted literature data into semantic triples.
• Employed knowledge graph embedding and semantic reasoning to uncover novel drug mechanisms and identify potential targets and drug candidates.

Main Results:

• The CovKG contains over 17 million triples, integrating diverse biomedical information.
• Semantic reasoning identified 41 novel drug mechanisms of action.
• The study identified 33 potential targets and 18 drug candidates, including 7 novel drugs and 3 key targets (ACE2, TMPRSS2, M protein).

Conclusions:

• Demonstrated the efficacy of a knowledge graph approach for coronavirus target discovery and drug repurposing.
• The methodology provides a framework for biomedical knowledge discovery applicable to other viruses and diseases.