Deep Learning in Virtual Screening: Recent Applications and Developments.

Area of Science:

• Computational Chemistry
• Medicinal Chemistry
• Bioinformatics

Background:

• Drug discovery is a lengthy and expensive process.
• Computational methods, including virtual screening, aid in designing new compounds.
• Machine learning (ML) has a long history in computer-aided drug discovery.

Purpose of the Study:

• To review recent advancements in ML, particularly deep learning (DL), for virtual screening.
• To explore DL applications in rational active compound discovery.
• To discuss current challenges and future directions in the field.

Main Methods:

• Review of ML and DL techniques for virtual screening.
• Discussion of compound and protein encoding strategies.
• Analysis of commonly used bioactivity and benchmark datasets.

Main Results:

• Deep learning significantly impacts rational active compound discovery.
• Various DL architectures and encoding methods are applicable.
• Availability of substantial chemical and bioactivity data fuels DL progress.

Conclusions:

• Deep learning offers powerful tools for accelerating drug discovery.
• Further research is needed to address current challenges and emerging problems.
• The integration of DL in virtual screening is crucial for future drug design.