Computational prediction of drug toxicity: the case of mutagenicity and carcinogenicity

Area of Science:

• Pharmacology
• Toxicology
• Computational Chemistry

Background:

• Pharmaceuticals are removed from the market due to toxicity.
• Early detection of drug toxicity is urgently needed.
• Predicting mutagenicity and carcinogenicity is vital for drug safety.

Purpose of the Study:

• To summarize methods for predicting mutagenicity and carcinogenicity from chemical structures.
• To identify limitations in current commercial prediction systems.
• To emphasize the need for integrating advanced chemical knowledge in toxicity prediction.

Main Methods:

• Review of structure-activity relationship (SAR) studies.
• Analysis of existing computational toxicology tools.
• Discussion of knowledge-based approaches in chemical safety assessment.

Main Results:

• Current commercial systems for predicting mutagenicity and carcinogenicity have limitations.
• Predicting complex toxicological endpoints solely from structure remains challenging.
• There is a gap between available computational tools and the need for accurate toxicity prediction.

Conclusions:

• Better approaches for early detection of drug toxicity are urgently required.
• Existing systems for predicting mutagenicity and carcinogenicity need improvement.
• Integrating higher levels of chemical knowledge is essential for advancing computational toxicology.