Advances in machine intelligence-driven virtual screening approaches for big-data

Area of Science:

• Computational chemistry and cheminformatics
• Drug discovery and development
• Artificial intelligence in medicine

Background:

• Virtual screening (VS) is a cornerstone of modern drug discovery, traditionally divided into ligand-based (LB) and structure-based (SB) methods.
• The increasing volume of chemical and biological data necessitates advanced computational approaches for efficient hit molecule identification.
• Machine intelligence (MI) offers powerful tools to enhance the speed and accuracy of VS, reducing time and resource consumption.

Purpose of the Study:

• To review and categorize various VS techniques, highlighting the integration of machine intelligence.
• To detail the implementation of machine learning and modern AI approaches in VS.
• To discuss the limitations of current VS methods and explore future prospects.

Main Methods:

• Review of literature on traditional LB and SB VS approaches.
• Analysis of machine learning and AI algorithm applications in VS.
• Categorization of integrated LB/SB techniques and big data-driven VS strategies.

Main Results:

• MI-driven VS enables rapid screening of ultra-large libraries, significantly reducing hit identification time.
• Integrated LB/SB approaches enhance prediction accuracy by considering both ligand and target properties.
• Advanced AI methods are crucial for handling big data and minimizing false positives in drug discovery.

Conclusions:

• Machine intelligence is revolutionizing virtual screening, making it a more progressive technology in drug discovery.
• The future of VS lies in advanced, intelligent solutions capable of managing big data and optimizing hit/lead identification.
• Continued development in AI and computational architecture is essential for overcoming current limitations and improving VS efficacy.