Multi-Objective Drug Molecule Optimization Based on Tanimoto Crowding Distance and Acceptance Probability.

Area of Science:

• Computational Chemistry
• Drug Discovery
• Bioinformatics

Background:

• Traditional molecular optimization methods are data-intensive and computationally expensive.
• Conventional genetic algorithms often yield similar solutions, limiting chemical space exploration and risking local optima.
• Existing approaches face challenges in maintaining molecular diversity during optimization.

Purpose of the Study:

• To present an improved genetic algorithm, MoGA-TA, for multi-objective drug molecular optimization.
• To enhance the exploration of chemical space and maintain population diversity in molecular optimization.
• To overcome limitations of traditional methods in data dependency and computational cost.

Main Methods:

• Developed MoGA-TA utilizing Tanimoto similarity-based crowding distance calculation.
• Implemented a dynamic acceptance probability population update strategy for evolutionary balance.
• Employed a decoupled crossover and mutation strategy for optimized molecular design.

Main Results:

• MoGA-TA demonstrated superior performance in drug molecule optimization compared to existing methods.
• The algorithm significantly improved the efficiency and success rate of molecular optimization.
• Evaluated effectiveness using metrics including success rate, dominating hypervolume, and internal similarity.

Conclusions:

• MoGA-TA is an effective and reliable method for multi-objective molecular optimization.
• The proposed approach enhances search space exploration and prevents premature convergence.
• This algorithm offers a promising solution for complex drug discovery challenges.