Innovative Multistage ML-QSAR Models for Malaria: From Data to Discovery.

Area of Science:

• Drug discovery and development
• Computational chemistry
• Parasitology

Background:

• Malaria remains a major global health threat, with nearly 247 million annual cases worldwide.
• Increasing Plasmodium parasite resistance to current antimalarial drugs necessitates novel therapeutic agents.

Purpose of the Study:

• To utilize artificial intelligence (AI) for the development of novel antimalarial compounds.
• To create predictive models for identifying compounds effective against multiple Plasmodium parasite life stages.

Main Methods:

• Development of multistage Machine Learning Quantitative Structure-Activity Relationship (ML-QSAR) models.
• Analysis of large chemical datasets to predict antimalarial activity.
• Experimental evaluation of selected compounds.
• Application of explainable AI (XAI) for molecular feature analysis.

Main Results:

• Six out of 16 evaluated compounds demonstrated dual-stage inhibitory activity against Plasmodium parasites.
• One compound exhibited inhibitory activity across all tested parasite life cycle stages.
• Explainable AI (XAI) provided insights into key molecular features driving compound efficacy.

Conclusions:

• AI-driven predictive modeling accelerates the identification and optimization of potential antiplasmodial compounds.
• The study highlights the potential of ML-QSAR and XAI in combating drug-resistant malaria.
• Novel compounds targeting multiple parasite stages offer a promising strategy against malaria.