AI-generated from publicly available materials.DeepCyte has introduced the DeeTox Atlas, a pioneering single-cell metabolomic reference designed to elucidate drug toxicity mechanisms. This initiative is set to engage in pilot programs with leading pharmaceutical firms in the near future.
The DeeTox Atlas serves as a foundational dataset derived from two separate single-cell metabolomics studies, encompassing around 100 toxic compounds, 300,000 individual cells, and approximately 500 metabolites per cell. The data includes over 3,000 single-cell measurements for each compound, organized according to a comprehensive four-tiered framework of toxicity mechanisms aligned with established Adverse Outcome Pathways (AOPs).
Instead of relying on new laboratory data for every compound, DeepCyte leverages a foundation model that predicts toxicity mechanisms for compounds not previously measured. This approach enhances scalability and minimizes reliance on traditional lab methods, as noted by co-founder and CEO Theodore Alexandrov. He emphasized that the DeeTox Atlas enables the identification of subtle molecular patterns associated with toxicity, which may be overlooked by conventional techniques lacking single-cell resolution.
As the DeeTox Atlas continues to grow with additional compounds and mechanisms, it aims to provide more actionable insights for toxicologists and safety scientists. This shift towards predictive, mechanism-based AI in toxicology could significantly transform the drug discovery process, allowing for earlier identification of safety concerns.