Machine Learning for Sequence and Structure-Based Protein-Ligand Interaction Prediction.

Area of Science:

• Computational chemistry and bioinformatics
• Drug discovery and development
• Machine learning in pharmacology

Background:

• Drug development is costly and time-consuming.
• Accurate prediction of drug-target interactions is crucial for efficient drug discovery.
• Machine learning (ML) approaches show promise in predicting protein-ligand interactions.

Purpose of the Study:

• To review computational methods for protein-ligand interaction prediction.
• To categorize and summarize ML models based on sequence and structure data.
• To discuss applications, evaluation, interpretability, challenges, and future directions in the field.

Main Methods:

• Overview of datasets used in protein-ligand interaction studies.
• Examination of various protein and ligand representation approaches.
• Classification of ML models (classical and deep learning) using sequence-based and structure-based criteria.

Main Results:

• Summarization of diverse ML models applied to protein-ligand interaction prediction.
• Discussion of evaluation metrics and model interpretability techniques.
• Exploration of the applications of these models in drug research.

Conclusions:

• Computational methods, particularly ML, offer a powerful approach to predict drug-target interactions.
• Further research is needed to address current challenges and advance future directions in the field.
• This review provides a comprehensive overview for researchers in drug discovery and computational biology.