A quantitative structure-toxicity relationships model for the dermal sensitization guinea pig maximization assay

Area of Science:

• Toxicology
• Computational Chemistry
• Dermatology

Background:

• Dermal sensitization is a significant health concern.
• Accurate prediction of skin sensitization is crucial for chemical safety assessment.
• Existing methods for assessing dermal sensitization can be resource-intensive.

Purpose of the Study:

• To develop quantitative structure-toxicity relationship (QSTR) models for predicting dermal sensitization.
• To classify chemical sensitizers into weak/moderate or severe categories.
• To ensure model predictions are confined to their domain of applicability.

Main Methods:

• Utilized results from the guinea pig maximization test (GPMT) on 315 chemicals.
• Developed two QSTR models: one for aromatics and one for aliphatics/one-benzene-ring compounds.
• Employed linear discriminant analysis with an optimum prediction space (OPS) algorithm for model development and validation.

Main Results:

• Achieved high cross-validated specificity (81–91%) and sensitivity (85–95%) for the QSTR models.
• Demonstrated good performance on an independent test set with 79% specificity and 82% sensitivity.
• The models successfully differentiate between weak/moderate and severe sensitizers.

Conclusions:

• The developed QSTR models provide a reliable computational approach for assessing dermal sensitization.
• These models can aid in prioritizing chemicals for further testing and risk assessment.
• The OPS algorithm ensures the robust application of QSTR models within their defined chemical space.