Deep Learning-driven research for drug discovery: Tackling Malaria.

Area of Science:

• Medicinal Chemistry
• Computational Biology
• Parasitology

Background:

• Malaria remains a significant global health burden, causing over 445,000 deaths annually.
• Emerging parasitic resistance to existing antimalarial drugs necessitates the urgent discovery of novel therapeutic agents.
• Current drug discovery pipelines are often slow and expensive, hindering the development of new treatments.

Purpose of the Study:

• To develop and apply deep learning-based Quantitative Structure-Activity Relationship (QSAR) models for predicting antimalarial activity and cytotoxicity.
• To accelerate the identification of novel antimalarial drug candidates through virtual screening.
• To experimentally validate computationally identified compounds against Plasmodium falciparum.

Main Methods:

• Development of binary and continuous QSAR models utilizing deep learning algorithms.
• Virtual screening of a large chemical compound database using the developed QSAR models.
• Experimental evaluation of top-ranked compounds against sensitive and multi-drug-resistant Plasmodium falciparum strains.

Main Results:

• Two compounds, LabMol-149 and LabMol-152, exhibited potent antiplasmodial activity at low nanomolar concentrations (EC50 <500 nM).
• These compounds demonstrated low cytotoxicity against mammalian cells, indicating a favorable safety profile.
• The deep learning approach successfully identified novel antimalarial agents meeting key development criteria.

Conclusions:

• Deep learning-powered QSAR modeling provides an efficient strategy for discovering new antimalarial drug candidates.
• LabMol-149 and LabMol-152 represent promising leads for next-generation antimalarial therapies.
• This computational approach can significantly expedite the drug discovery process for neglected diseases like malaria.