An introduction to GitHub and its significance for AI-driven drug discovery

Area of Science:

• Computational chemistry and cheminformatics
• Artificial intelligence in pharmacology
• Bioinformatics and computational biology

Background:

• GitHub is integral to AI-driven drug discovery, supporting code sharing and reproducible research.
• The increasing reliance on GitHub necessitates robust methods for repository evaluation.
• Researchers require guidance on identifying, assessing, and reusing AI drug discovery tools hosted on GitHub.

Purpose of the Study:

• To summarize key GitHub concepts relevant to AI drug discovery research.
• To propose a practical framework for evaluating GitHub repositories in this field.
• To analyze trends and topic density of drug discovery repositories on GitHub.

Main Methods:

• Summarized essential GitHub features for research (repositories, documentation, licensing, etc.).
• Developed a framework for navigating and assessing drug discovery repositories.
• Conducted a Scopus trend analysis (2013-2024) and keyword-based GitHub repository counts.

Main Results:

• GitHub repositories are vital research outputs in AI drug discovery.
• Essential criteria for repository evaluation include documentation, reproducibility (pinned environments/containers), licensing, and automated testing.
• Enhanced validation and governance are needed for reliable translational and industrial application.

Conclusions:

• GitHub repositories should be treated as peer-assessable research outputs.
• Implementing clear documentation, reproducibility measures, licensing, and testing is fundamental.
• Further validation and governance are critical for advancing AI drug discovery from exploration to industrial application.