Deep Origin Secures $31.7M ARPA-H Contract to Build FDA-Qualifiable Virtual Human Platform for Drug Testing
核心洞察
Deep Origin (搜索), founded in 2022, secured a $31.7 million contract from ARPA-H to develop an FDA-qualifiable in silico prediction platform aimed at reducing and ultimately replacing animal testing.
The company's computational platform contributed to a peer-reviewed Cell study where simulations helped identify a compound that activates cell-death programs in diffuse large B-cell lymphoma (搜索).
Over 90% of drugs fail during clinical development due to poor predictivity of preclinical models, a gap Deep Origin (搜索) addresses by layering safety predictions onto drug candidates before they enter the clinic.
Deep Origin (搜索), a computational drug discovery company founded in 2022, has secured a $31.7 million contract from the Advanced Research Projects Agency for Health (搜索) (ARPA-H) to develop an FDA-qualifiable in silico prediction platform designed to reduce and ultimately replace animal testing. The funding, awarded through ARPA-H's Catalyst program, marks a significant milestone for the company's ambitious "virtual human" initiative, which combines molecular physics with artificial intelligence to create predictive models that could fundamentally reshape preclinical drug development.
The contract arrives against a backdrop of stark industry statistics: over 90% of drugs fail during clinical development, largely because preclinical models poorly predict how compounds will behave in humans. "What we're after is the quality of the prediction itself. That's the predictivity crisis, a systematic gap between what works in preclinical models and what happens in the clinic," said Natalie Ma, chief business officer and co-founder of Deep Origin (搜索), in an interview with PharmaVoice.
A Real-World Validation in Lymphoma Research
Deep Origin (搜索)'s technology recently demonstrated its potential in a peer-reviewed study published in Cell, where the company's docking and molecular dynamics simulations contributed to identifying a compound that activates cell-death programs in diffuse large B-cell lymphoma (搜索). "Deep Origin's computational simulations flagged the compound that would kill cancer cells most effectively — matching results achieved at the bench," said Garegin Papoian, Ph.D., co-founder and chief scientific officer of Deep Origin. "This is a predictivity goal of in silico drug discovery — to determine the candidates most likely to achieve desired results prior to wet lab experimentation. This was a meaningful real-world test of our systems."
Building the Virtual Human, One Organ at a Time
The company's approach distinguishes itself by tying models together across biological scales — from quantum mechanical interactions and docking, through protein dynamics and cell-state changes, up to organ-level exposure and whole-body physiology. This multi-scale architecture allows safety questions to be traced back to concrete mechanistic explanations, a feature Ma contrasts with "black-box" machine learning predictors and foundational models that lack explainability.
Ma confirmed that Deep Origin (搜索) plans to bring its first organ models online later this year or early next year. "The goal is to produce more accurate, human-relevant safety predictions before a molecule advances, bringing organ-level models online, tracing predicted drug-induced liver injury or tox signals back to specific off-target interactions and using those insights to redesign or de-prioritize risky compounds," she said.
A Measured Approach to a Transformative Vision
Despite the long-term ambition of running in silico clinical trials where "clinical trials are a formality," Ma emphasized that the company's immediate focus is pragmatic. "We don't position our immediate work as 'replacing' clinical trials. Right now, it's about reducing late safety failures and animal use, and giving programs more confidence going into the clinic," she said.
The company works with partners through three engagement models: discovery partnerships that apply its platform and scientists to partners' most challenging drug discovery programs, direct SaaS access to its computational platform, and co-development of Deep Origin (搜索)-originated programs. "A lot of the value comes from being able to increase success rate in the lab while still finding high-quality hits and then layering safety predictions on top, so what moves forward has a much better chance of working in humans, not just in animals," Ma explained.
Looking further ahead, Ma envisions researchers inputting a small-molecule structure and dose into virtual humans parameterized for different genetic backgrounds and comorbidities, allowing scientists to predict outcomes across multiple organs. However, she cautioned that realizing this vision "will require extensive validation and ongoing dialogue with the FDA," describing it as "a long regulatory horizon, gated by validation and dialogue with the FDA."
The virtual human framework, as Ma described it, is distinct from "digital twin" approaches that typically rely on statistical averages of patient parameters. Instead, Deep Origin (搜索)'s system models human biology across multiple organ systems, incorporating representations of biology across scales — from gene and protein interactions to macromolecule and small-molecule dynamics — to capture the nuance of individual patients.