DeepCyte has launched DeeTox Atlas, a proprietary single-cell metabolomic reference atlas designed to help pharmaceutical companies predict drug toxicity mechanisms earlier in the discovery process. The company said the platform is the world’s first single-cell metabolomic reference atlas focused on drug toxicity mechanisms.
DeeTox Atlas is built from two independent single-cell metabolomics perturbation studies spanning approximately 100 toxicant compounds and 300,000 cells.
The dataset captures approximately 500 metabolites per cell and more than 3,000 single-cell measurements for each compound across six biological replicates.
Each compound is mapped to a curated four-level hierarchy of toxicity mechanisms tied to established Adverse Outcome Pathways.
DeepCyte is using the dataset to train AI models that can predict toxicity mechanisms for compounds that have not previously been measured by the platform.
The approach is intended to reduce the need to generate new wet-lab data for every individual compound.
As the atlas adds additional compounds and mechanisms, DeepCyte expects the expanding dataset to improve the scalability and predictive capabilities of its models.
The company’s underlying thesis is that single-cell measurements can reveal biological signals that may be missed when researchers examine averaged measurements across larger cell populations.
DeepCyte said its validation studies have identified toxicity-related molecular patterns within small cell subpopulations that can be difficult to detect without single-cell resolution.
Those patterns can then become inputs for models designed to predict and explain potential safety liabilities.
The goal is to shift more toxicology analysis toward mechanism-based prediction earlier in drug discovery, potentially giving researchers additional information before compounds advance into more expensive stages of development.
DeepCyte plans to expand DeeTox Atlas with additional compounds, toxicity mechanisms and biochemical and clinical information.
The company believes this could make its predictions increasingly useful to toxicologists, medicinal chemists and safety scientists evaluating drug candidates.
Commercial validation is also moving forward.
DeepCyte said enterprise pilot programs with global pharmaceutical companies are expected to begin in the coming months.
The company is currently advancing pilots with three of the world’s top 20 pharmaceutical companies.
DeepCyte has also appointed Mona Lakkis as Senior Vice President of Partnerships to lead enterprise commercialization.
Lakkis previously served as Senior Vice President and General Manager at Owkin.
Her appointment comes as DeepCyte moves from building its proprietary toxicity dataset toward broader engagement with pharmaceutical companies seeking AI-based tools for preclinical safety assessment.
DeepCyte is positioning DeeTox Atlas as a foundation dataset rather than simply a database of previously observed toxicity results.
By training models across mechanistically categorized single-cell data, the company aims to predict how previously untested compounds could trigger biological pathways associated with toxicity.
That strategy could become increasingly valuable if the system can identify potential safety problems earlier while explaining the biological mechanisms behind its predictions.
KEY QUOTES:
“AI in toxicology is only as good as the biological data it learns from. DeeTox Atlas lets us find subtle molecular patterns tied to key toxicity mechanisms, patterns our validation studies show are expressed in small subpopulations of cells and are effectively invisible to methods lacking single-cell resolution, and turn them into predictive models.”
“Our vision is to move toxicology from reactive laboratory testing toward predictive, mechanism-based AI that surfaces and explains safety liabilities earlier in drug discovery.”
Theodore Alexandrov, Co-Founder and CEO of DeepCyte
