Development of an expert system rulebase for identifying contact allergens

Area of Science:

• Toxicology
• Computational Chemistry
• Dermatology

Background:

• Currently, no in vitro methods exist for identifying skin sensitizers (contact allergens).
• Expert systems can integrate chemical structure-toxicity knowledge.
• A historical database of 294 substances tested for skin sensitization was available.

Purpose of the Study:

• To derive structural alerts for skin sensitization.
• To develop an expert system for predicting skin sensitization.
• To potentially reduce animal testing in toxicology.

Main Methods:

• Utilized a historical database of 294 substances tested in the guinea pig maximization test.
• Grouped substances by likely reaction mechanisms with skin proteins.
• Derived empirical structural alerts for molecules with similar chemical functionality.
• Incorporated 40 structure-activity rules into the DEREK expert system.

Main Results:

• Developed 40 structure-activity rules for skin sensitization.
• Integrated these rules into the DEREK expert system.
• Demonstrated a potential method for preliminary toxicological hazard identification.

Conclusions:

• The developed expert system and rulebases show potential for preliminary screening of skin sensitizers.
• This in silico approach may lead to a reduction in animal use for toxicological assessments.
• Further development could enhance the identification of contact allergens.