A beginner's approach to deep learning applied to VS and MD techniques.

Area of Science:

• Computational Chemistry
• Medicinal Chemistry
• Biochemistry
• Machine Learning

Background:

• Computational chemistry and molecular modeling are integral to modern drug development.
• In silico methods, including virtual screening (VS) and molecular dynamics (MD) simulations, are indispensable in drug discovery and biophysical studies.
• Existing VS and MD methods face limitations in hardware and algorithms.

Purpose of the Study:

• To provide a comprehensive review of recent deep learning (DL) applications in molecular modeling.
• To guide computational chemists in integrating DL into their research.
• To consolidate scattered knowledge on DL in molecular modeling.

Main Methods:

• The review surveys recent applications of DL across various molecular modeling workflows.
• Sections are organized by DL integration points: improving VS, enhancing MD simulations, aiding interatomic force calculations, and analyzing MD trajectories.

Main Results:

• Deep learning offers solutions to overcome limitations in traditional VS and MD methods.
• DL can lead to more accurate and efficient molecular modeling results.
• DL applications expedite data analysis of simulation results.

Conclusions:

• Deep learning has the potential to revolutionize molecular modeling techniques.
• Integrating DL can significantly improve the efficiency and accuracy of drug discovery and biophysical research.
• Further exploration and adoption of DL are encouraged for computational chemists.