AI-Enabled Pharmacokinetics & PBPK Modeling
Predicting how drugs move through the body—long before clinical trials begin.
Most drug failures stem from how a therapy behaves once inside the body—absorption, distribution, metabolism, excretion, and toxicity (ADMET). Getting this wrong leads to costly failures and patient risk. Over 90% of candidates fail primarily due to poor disposition/safety, not target binding.
Existing tools often fall short for complex biologics (peptides, antibodies) and can’t reliably forecast in-body performance or drug–drug interactions. AI has transformed molecule design, but robust AI for human disposition prediction remains a critical gap.
AI-enabled physiologically based pharmacokinetic (PBPK) models that integrate pharmacometrics and quantitative systems pharmacology to simulate behavior across populations and reduce reliance on animal testing.
Fewer failed trials, faster go/no-go decisions, more precise dosing—plus earlier, safer integration of disposition predictions in the development process.
AI-Enabled Peptide Design
World‑leading expertise in peptide design to unlock programmable, next‑generation therapies.
Peptides offer specificity and safety advantages and can be tailored to disease pathways with precision.
Stability, rapid degradation, and hard-to-predict in-body performance have historically slowed progress.
I²D³ leverages UW-built AI diffusion models (Bhardwaj Lab) that have shown strong accuracy in peptide target-binding prediction, paired with robust in-house datasets (including oral bioavailability and metabolic stability) to accelerate design and reduce trial-and-error.
Faster design cycles, reduced R&D waste, safer and more effective drugs—and a data/AI platform that scales to additional biologics after peptides and antibodies.
AI-Enabled Antibody Engineering
Engineering antibodies to expand the power of biologics and targeted therapies.
Antibodies are central to cancer, infectious disease, and chronic-condition therapies—and essential to precision medicine.
Specificity, immunogenicity, durability, and manufacturability remain complex and slow the pipeline.
AI-informed antibody engineering integrated with predictive modeling and developability assessment—first applied to peptides and antibodies, then extended to other biologics—to improve targeting, stability, and translational reliability.
More effective biologics, faster translation, and a path to IP, licensing, and spin-outs as tools mature.