Draper and Partners Awarded ARPA-H Funding for In Silico Drug Development Models

The $26.7 million award will support the development of new computer models that mimic human biology to predict drug safety and effectiveness.

Published on Feb. 24, 2026

Got story updates? Submit your updates here. ›

Draper, a non-profit research and development company, has received a $26.7 million award from the Advanced Research Projects Agency for Health (ARPA-H) to develop a new generation of in silico models of human physiology. The models will leverage artificial intelligence and machine learning to predict the safety and effectiveness of drug candidates, with the goal of ensuring only the most promising and safest medicines move forward to patients. Draper's partners in the project include Revalia Bio, the Krishnaswamy Lab at Yale University, and LifeShare Network.

Why it matters

Modern drug development and approval pathways require advanced methods for accurately predicting safety profiles. The new in silico models aim to overcome the limitations of existing human-based models by integrating data from multiple sources, including electronic medical records, donated human organs, tissue biopsies, and Draper's own microphysiological systems. These models could serve as a direct replacement for animal models, enabling faster and cheaper drug development.

The details

The project will develop an integrated 'Human Data Stack' that combines four layers of human data: patient data from electronic medical records, macrophysiological data from donated human organs, cellular and molecular data from tissue biopsies, and microphysiological data from Draper's PREDICT96 system that evaluates human cells in a complex environment. This data will feed continuous learning models that leverage AI and machine learning to support patient-specific predictions of absorption, distribution, metabolism, excretion, and toxicity (ADME-Tox).

  • Draper recently received the $26.7 million award from ARPA-H.
  • The research will initially focus on the human liver and kidney as proof of concept.

The players

Draper

A non-profit research, development, and manufacturing company that solves important challenges, with a focus on strategic systems, space systems, electronic systems, and biotechnology systems.

Revalia Bio, Inc.

A company that replaces animal models with scalable, AI-powered 'Human Data Trials' built on living human organs, aiming to collapse drug development timelines and costs.

Krishnaswamy Lab at Yale University

A research lab that develops data geometric, topological, dynamic, and deep learning methods for analyzing and representing biomedical data.

LifeShare Network

A non-profit organization that fosters a commitment to serving donor heroes and their families, and is at the forefront of transplant research, innovation, and education.

Greg Tietjen

CEO of Revalia Bio.

Katie Hulse

Distinguished Member of the Technical Staff at Draper.

Got photos? Submit your photos here. ›

What they’re saying

“Collectively, we believe an integrated Human Data Stack can be leveraged to develop a new class of Human Data Trial that can, in the near term, serve as a direct replacement to animal models under the current paradigm of drug development.”

— Greg Tietjen, CEO, Revalia Bio

“A unique aspect of the integrated Human Data Stack is the blending of static, end point data—patient and biopsy data—with two separate dynamic experimental systems—the macro- and microphysiological data. In these layers, dynamic human experiments can be run to rapidly test hypotheses and refine our in silico models without risk to living patients.”

— Greg Tietjen, CEO, Revalia Bio

“As part of an integrated Human Data Stack, Draper's PREDICT96 system is able to provide a crucial layer of human data to support direct replacement of animal models, enabling faster and cheaper drug and medical countermeasure development.”

— Katie Hulse, Distinguished Member of the Technical Staff, Draper

What’s next

Subsequent research, if funded, will involve using the in silico models developed in Phase 1 towards generating predictive ADME-Tox data for use in Investigational New Drug (IND) filings with the FDA.

The takeaway

This project represents a significant step forward in the development of advanced in silico models that can accurately predict drug safety and effectiveness, potentially reducing reliance on animal models and accelerating the drug development process. The integration of diverse human data sources, coupled with the latest AI and machine learning techniques, could lead to a new paradigm in how we approach drug discovery and approval.