Cedars-Sinai KronosRx team awarded $5M to advance human-relevant drug toxicity prediction

25 September 2025

Cedars-Sinai Medical Center based in Los Angeles has been awarded $5M by Advanced Research Projects Agency for Health (ARPA-H) to develop a human-based AI-driven KronosRx platform for improved prediction of adverse drug effects.

The ARPA-H Computational ADME-Tox and Physiology Analysis for Safer Therapeutics (CATALYST) program aims to revolutionize preclinical drug safety prediction by facilitating the development of next gen in vitro and in silico models that accurately capture human-specific and patient-specific physiology. By employing latest advances in complex in vitro, AI and computer models, CATALYST ambitions to derisk drug development, ameliorate safety of therapeutic candidates and accelerate precision medicine.

To succeed in this mission, Cedars-Sinai has leveraged its diversity of skills to build a team that will collaborate across disciplines:

• Nicholas Tatonetti, PhD, the project’s lead investigator and vice chair of Computational Biomedicine at Cedars-Sinai

• Clive Svendsen, PhD, executive director of the Cedars-Sinai Board of Governors Regenerative Medicine Institute

• Arun Sharma, PhD, director of the Cedars-Sinai Center for Space Medicine Research in the Board of Governors Regenerative Medicine Institute

• Graciela Gonzalez-Hernandez, PhD, vice chair for Research and Education in the Department of Computational Biomedicine

Dr. Tatonetti is passionate about the integration of real-world and high-dimensional multi-omics patient data into computational methods of identification of adverse drug effects and drug-drug interactions. Dr. Svendsen team advances biomedical research into human neurodegenerative disorders by employing organ-on-chip devices that capture human-relevant disease pathophysiology. Dr. Sharma lab employs human-derived and patient-specific organoids and heart-on-chip systems to assess drug-induced cardiotoxicity and to investigate the mechanisms of human cardiovascular disorders. Dr. Gonzalez-Hernandez specializes on the extraction of unstructured real-world patient data, such as electronic health records data, using natural language processing and machine learning.

“Each year, many promising drugs fail in trials because animal tests and short-term lab studies cannot predict how medicines behave in real people over time,” said Nicholas Tatonetti, PhD, the KronosRx project lead investigator. “These failures delay lifesaving treatments and drive up drug development costs.”

Very poor negative predictivity of animal testing is a consequence of myriad human-specific features of anatomy and physiology  Clark & Steger-Hartmann, Reg. Toxicol. and Pharmac., Jul 2018. In certain cases, vulnerability to drug-induced toxicity may arise from an interplay of drug characteristics and patient-specific risk factors - age, gender, comorbidity, polypharmacy, immune status, lifestyle, genetic polymorphism. However, the individual and combined effects of these risk factors cannot be recapitulated in animals, resulting in drug-induced morbidity and mortality.

To advance human-relevant drug toxicity prediction, the Cedars-Sinai KronosRx team will therefore employ its biobank of patient iPSC-derived organoids that capture human-specific molecular, histological and functional features.  The AI models will be trained using millions of anonymous real-world patient data from Cedars-Sinai’s extensive electronic health record network. The resulting platform will be able to predict patient responses to drugs over time and across the diverse population of patients reflected in the Cedars-Sinai data.