Deep Origin Launches In Silico Drug Safety Prediction Platform
News related to:Deep Origin · 2 min read
SOUTH SAN FRANCISCO, Calif., Sept. 24, 2026 /CourierPR/ -- Deep Origin, a company developing computational discovery systems, has launched a groundbreaking in silico drug safety prediction platform. The platform, part of the PREDICTS consortium, is funded by a $31.7 million Other Transaction Agreement (OTA) from the Advanced Research Projects Agency for Health (ARPA-H) and is aimed at revolutionizing drug safety prediction.
The consortium, led by Deep Origin, includes ImmVue Therapeutics, Synko, Sanford Burnham Prebys, and SyzOnc. These organizations are now using Deep Origin’s ADMET (absorption, distribution, metabolism, excretion, toxicity) models to assess and filter millions of compounds, prioritizing drug candidates for further development. The ADMET-NOW platform comprises 77 machine learning-based models, with 62 now available for pilot participants.
These models are built on Togo, Deep Origin’s foundational chemistry model, which can train new property models on fewer than 1,000 data points. According to the company, 89% of the models outperform the best-performing models identified in the literature, including genotoxicity, human ether-à-go-go-related gene (hERG) inhibition, and drug-induced liver injury (DILI).
The PREDICTS consortium is part of the CATALYST program, which aims to develop human-based models to predict drug safety and toxicity more accurately. The program is led by Michael Patterson and is funded by ARPA-H. The goal is to enable safer and faster drug development, particularly for rare disease populations, and to meet the targets of the U.S. Food and Drug Administration’s (FDA) Modernization Act.
Deep Origin’s models are not just about predicting safety and toxicity; they are also about improving the overall drug development process. By using these in silico models, drug developers can save time and resources that would otherwise be spent on extensive wet-lab testing. The models will be continuously improved as the participating organizations send their lab results back to extend the chemical space the models are trained on.
The company is also developing Virtual Human Avatars of Toxicology (VHAT) to increase the human-relevance of safety predictions in preclinical development. These virtual humans will simulate the physical and chemical events inside a body to predict drug safety, including PK and permeability, liver metabolism, and the interaction of drugs with proteins and cells.
Deep Origin, co-founded by Michael Antonov and Garegin Papoian, has been backed by more than $50 million in capital and over $32 million in non-dilutive funding, including the ARPA-H CATALYST award. The company’s goal is to close the gap between preclinical predictions and clinical outcomes, making drug development more efficient and effective.
The in silico models developed by Deep Origin are part of a broader effort to reduce reliance on animal testing in preclinical drug development. By leveraging advanced AI and computational models, the company aims to provide a more accurate and humane approach to drug safety assessment.