Predictive modeling of antibacterial activity of ionic liquids by machine learning methods

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Materials Science

Background:

• Ionic liquids (ILs) possess tunable structures and bioactivity, making them candidates for novel biocide development.
• Computational methods offer efficient, cost-effective alternatives to experimental synthesis for evaluating potential antibacterial agents.
• Predicting the minimal inhibitory concentration (MIC) of compounds is crucial for biocide development.

Purpose of the Study:

• To develop quantitative structure-activity relationship (QSAR) models for predicting the antibacterial activity of ionic liquids (ILs).
• To identify key structural descriptors influencing the minimal inhibitory concentration (MIC) of ILs against specific bacterial pathogens.
• To validate the predictive models using newly synthesized and purchased ionic liquids.

Main Methods:

• Quantitative structure-activity relationship (QSAR) modeling using over 800 data points.
• Random forest algorithm applied with AlvaDesc molecular descriptors.
• SHapley Additive exPlanations (SHAP) method for model interpretability.
• In vitro testing of synthesized and purchased ionic liquids.

Main Results:

• The random forest model, utilizing AlvaDesc descriptors, demonstrated superior predictive performance for antibacterial activity against Staphylococcus aureus, Escherichia coli, and Pseudomonas aeruginosa.
• The SHAP method identified significant descriptors influencing the minimal inhibitory concentration (MIC) of ILs.
• Newly synthesized amino acid ILs and purchased halogenide ILs validated the developed QSAR models.

Conclusions:

• QSAR modeling, particularly with random forest and AlvaDesc descriptors, is a reliable method for predicting the antibacterial efficacy of ionic liquids.
• The study provides a validated computational tool for the rational design of novel ionic liquid-based biocides.
• The developed models and data are publicly accessible for further research and application.