Joint molecular optimization for efficient virtual screening via machine learning

Abstract

Joint molecular optimization serves as the central focus of this research, which explores the synergistic integration of machine learning (ML) and virtual screening (VS) for drug discovery. This study addresses these limitations by integrating molecular docking and ML to predict and balance multiple molecular properties concurrently. Traditional VS methods often fall short by primarily targeting single molecular properties, such as binding affinity, overlooking the necessity of simultaneously optimizing multiple crucial parameters like ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) profiles. Key objectives include developing and evaluating multi-objective optimization strategies and employing ML models (e.g., logistic regression, SVM, neural networks) to enhance predictive accuracy and computational efficiency. This research balances efficacy and safety metrics to refine VS and demonstrates the effect of combining these strategies on benchmark data. This work contributes a controlled, reproducible framework and an explicitly defined multi-objective prioritization procedure intended to support more efficient and transparent virtual screening, offering valuable insights for researchers in drug discovery, computational chemistry, and ML. Claims of reduced clinical attrition or improved clinical success would require prospective experimental validation and lie beyond the scope of this computational study.