An Analysis of QSAR Research Based on Machine Learning Concepts.

Area of Science:

• Computational Chemistry
• cheminformatics
• Drug Discovery

Background:

• Quantitative Structure-Activity Relationship (QSAR) models link chemical structures to biological activities.
• Machine learning (ML) techniques are increasingly vital for developing accurate QSAR models.
• Existing literature lacks a recent, comprehensive analysis of ML algorithms in QSAR.

Purpose of the Study:

• To systematically review machine learning applications in QSAR modeling.
• To introduce the ML-QSAR analytical framework for future research.
• To facilitate algorithm selection, method development, and comparative analysis of QSAR methodologies.

Main Methods:

• A structured categorization of QSAR research based on machine learning models.
• Introduction of assessment criteria for evaluating QSAR models.
• Qualitative analysis of machine learning algorithms in QSAR.

Main Results:

• The ML-QSAR framework provides a structured approach to QSAR research.
• The framework categorizes ML algorithms and introduces assessment criteria.
• A qualitative analysis guides the selection and development of QSAR methods.

Conclusions:

• The ML-QSAR framework offers a valuable resource for researchers in the field.
• It supports informed decision-making for selecting and improving QSAR modeling strategies.
• This systematic review and framework enhance the comparative study of ML-based QSAR methodologies.