CORAL: Models of toxicity of binary mixtures

Quantitative structure–activity relationships (QSAR) for toxicity of binary mixtures (expressed as pEC50 (i.e. log[1/EC50], logarithm of the inverse of the effective concentration required to bring about a 50% decrease in light emission), for Photobacterium phosphoreum) have been developed. The simplified molecular input-line entry system (SMILES) was used as the representation of the molecular structure of components of binary mixtures. Using the Monte Carlo technique the SMILES-based optimal descriptors were calculated. One-variable correlations between the optimal descriptors and toxicity of the binary mixtures were analyzed to develop a predictive model. Six random splits of the data into sub-training, calibration, and test sets were tested. A satisfactory statistical quality of the model was achieved for each above-mentioned split.

► The quantitative model for toxicity of binary mixtures is suggested. ► The molecular structure is represented by SMILES. ► The model of toxicity is built up by the CORAL software available on the Internet.