DeepCyte Launches AI-Powered Atlas for Drug Toxicity Prediction

DeepCyte Inc. has launched DeeTox Atlas™, a single-cell metabolomics reference atlas designed to advance AI-based prediction of drug toxicity mechanisms. 

The platform aims to help pharmaceutical companies identify potential safety risks earlier in the drug development process, with enterprise pilot programmes expected to begin in the coming months.

The atlas combines data from two independent single-cell metabolomics studies covering around 100 toxicant compounds, 300,000 cells and approximately 500 metabolites per cell. 

Each compound is supported by more than 3,000 single-cell measurements across six biological replicates and mapped to a four-level toxicity mechanism hierarchy linked to established Adverse Outcome Pathways (AOPs).

DeepCyte is using the dataset to train a foundation model capable of predicting toxicity mechanisms for compounds that have not been directly measured. 

This approach could reduce reliance on repeated laboratory testing and support more scalable safety assessments.

The company plans to expand DeeTox Atlas with additional compounds, toxicity mechanisms, biochemical information and clinical data. 

The expanded dataset is expected to improve the platform's predictive capabilities and provide actionable insights for toxicologists, medicinal chemists and safety scientists.