PredAOT

PredAOT predicts acute oral toxicity of small chemical compounds in mice and rats to support early-stage drug development decision-making.

Key Features:

• Multiple Random Forest Models: Uses multiple random forest models as the core machine-learning approach for toxicity prediction.
• Training Data: Models trained on datasets comprising 6,226 compounds evaluated in mice and 6,238 compounds evaluated in rats.
• Dual-Species Prediction: Provides predictions for both mice and rats.
• Benchmarking: Performance has been benchmarked against existing tools, showing similar or superior predictive accuracy.
• Target Compounds: Focuses on small chemical compounds relevant to early-stage drug development.

Scientific Applications:

• Acute Oral Toxicity Screening (Mouse): Predicts acute oral toxicity outcomes for candidate compounds in mice.
• Acute Oral Toxicity Screening (Rat): Predicts acute oral toxicity outcomes for candidate compounds in rats.
• Early-Stage Drug Candidate Prioritization: Supports prioritization of small-molecule candidates based on predicted toxicity profiles.
• Cross-Species Assessment: Enables assessment across two rodent species to inform comparative toxicology decisions.

Methodology:

Multiple random forest models trained on datasets of 6,226 mouse compounds and 6,238 rat compounds.