Computational approaches for skin sensitization prediction.

Area of Science:

• Toxicology
• Computational Chemistry
• Dermatology

Background:

• Skin sensitization is a significant safety concern across industries, affecting consumers and workers.
• Traditional animal testing methods for assessing skin sensitization are increasingly being replaced by non-animal and in silico approaches.
• Predicting the sensitization potential and potency of chemicals is vital for risk assessment and regulatory compliance.

Purpose of the Study:

• To provide a comprehensive overview of computational models for skin sensitization prediction developed in the last 10 years.
• To discuss the scope, limitations, and validation of various computational approaches.
• To highlight accessible models and comparative performance assessments.

Main Methods:

• Review of rule-based approaches, read-across, and (quantitative) structure-activity relationship ((Q)SAR) modeling.
• Analysis of hybrid and combined computational methods.
• Integration of computational models with experimental data.

Main Results:

• A detailed examination of diverse computational models for predicting skin sensitization.
• Discussion on the accessibility and validation status of these models for the scientific community.
• Comparative performance assessments of different predictive approaches.

Conclusions:

• Computational models offer promising alternatives to animal testing for skin sensitization assessment.
• The review emphasizes the importance of model validation and accessibility for practical application.
• Understanding the strengths and limitations of various in silico methods is key for accurate risk assessment.