Ensemble-Based Modeling of Chemical Compounds with Antimalarial Activity

Area of Science:

• Computational chemistry
• Medicinal chemistry
• Drug discovery

Background:

• Malaria is a deadly tropical disease caused by Plasmodium parasites, transmitted by Anopheles mosquitoes.
• Existing antimalarial therapies exhibit high toxicity.
• Quantitative Structure-Activity Relationship (QSAR) studies are vital in drug design.

Purpose of the Study:

• To develop computational models for identifying antimalarial hit compounds.
• To leverage QSAR for efficient drug discovery pipelines.

Main Methods:

• Compiled and curated a dataset for antimalarial activity modeling.
• Evaluated various ensemble-based classification methods.
• Selected an ensemble combining Genetic Algorithms and Majority Vote for optimal performance.

Main Results:

• Ensemble modeling effectively addresses QSAR for heterogeneous antimalarial datasets.
• The chosen ensemble demonstrated superior virtual screening performance.
• Identified promising strategies for discovering novel antimalarial agents.

Conclusions:

• Ensembles utilizing Genetic Algorithms for base model selection and Majority Vote for aggregation are highly effective.
• This approach enhances the identification of potential antimalarial drug candidates.
• Ensemble modeling is a robust strategy for QSAR in antimalarial drug discovery.