DeepCyte launched DeeTox Atlas, a single-cell metabolomic reference atlas designed to map drug toxicity mechanisms at cellular resolution. The company said it plans enterprise pilot programs with global pharmaceutical partners to begin in coming months, positioning the dataset as a foundation for toxicity prediction and safety liability triage. DeepCyte describes DeeTox Atlas as built from two independent single-cell metabolomics perturbation studies spanning roughly 100 toxicant compounds, 300,000 cells, and about 500 metabolites per cell. Company materials cite around 3,000 single-cell measurements per compound across six biological replicates, with compounds mapped into a curated four-level hierarchy of toxicity mechanisms anchored to established adverse outcome pathways (AOPs). Rather than requiring new wet-lab data for every chemical, DeepCyte says it trained a foundation model to predict toxicity mechanisms for compounds it has not measured. Chief executive Theodore Alexandrov framed the platform as shifting toxicology toward predictive, mechanism-based safety screening—particularly because company validation studies indicate key signals can reside in small cell subpopulations. If adopted at scale, single-cell metabolomic “mechanism libraries” like DeeTox Atlas could become a differentiator in early drug discovery, helping safety teams prioritize chemotypes before liabilities emerge in later stages.