Can the Monte Carlo method predict the toxicity of binary mixtures?

Area of Science:

• Environmental toxicology
• Computational chemistry
• Ecotoxicology

Background:

• Ecosystem risk assessment often relies on single toxicant data, which is unrealistic as organisms are exposed to complex mixtures.
• Assessing the combined effects of toxicants in mixtures is a long-standing challenge in ecotoxicology.

Purpose of the Study:

• To develop and validate a Quantitative Structure-Activity Relationship (QSAR) model for predicting the joint toxicity of binary chemical mixtures.
• To utilize Simplified Molecular Input Line Entry System (SMILES) for molecular structure representation in QSAR modeling.

Main Methods:

• Employed QSAR modeling to predict joint toxic effects in binary mixtures.
• Utilized SMILES notation for representing molecular structures of mixture components.
• Incorporated Monte Carlo optimization with the Index of Ideality of Correlation (IIC) to refine 2D-optimal descriptors.

Main Results:

• The QSAR models demonstrated high predictive accuracy, validated by comparison with experimental data.
• The average statistical performance on the validation set included n=25, R²=0.95, and RMSE=0.375.
• The use of IIC in Monte Carlo optimization improved the SMILES-based QSAR models.

Conclusions:

• QSAR modeling, particularly when enhanced with SMILES and IIC optimization, provides a robust approach for assessing the ecotoxicological impact of chemical mixtures.
• The developed methodology offers a valuable tool for environmental risk assessment, moving beyond single-substance evaluations.