We did not discover UNDC-005 in a lab. We found it in the scientific record, where the human evidence already existed.
Two platforms, one lifecycle
Our artificial intelligence ecosystem comprises two interconnected platforms. Precision.AI is focused on novel target and drug discovery. DrugTX.AI is designed to optimize the drug development lifecycle, from clinical operations and manufacturing through commercialization.
Most AI drug development has concentrated on the earliest stages — target identification, molecular design, in-silico screening. We believe a significant share of clinical-stage failures are not caused by insufficient discovery science but by execution: trial design that does not optimize for regulatory acceptance, enrollment inefficiency, CMC gaps, and commercialization planning that begins too late. These are operational failures, not scientific ones, and they are largely unaddressed by the current generation of discovery platforms.
Drug development remains subject to substantial risk regardless of the tools employed.
Target and drug discovery
Target discovery
TargetAppy ingests multi-modal datasets, including proprietary patient data secured through European hospital partnerships, and applies multiparameter analysis to score candidate targets. The output is a curated list of targets confirmed in human samples.
Disease modeling and human genetics
Computational biostatistics re-analyze GWAS and multi-omics datasets to identify functionally related gene clusters implicated in disease. This methodology identified the phospholipid and sphingolipid dysregulation that forms the biological foundation for UNDC-005.
De novo drug design
Generative models and biophysics optimization tools design novel molecular structures, including antibodies and antibody-drug conjugates, optimized for specific biological properties.
Diagnostics
BPD-AI applies imaging analysis and multi-omics data to identify disease subtypes and predict bronchopulmonary dysplasia severity in neonates. Patent pending.
Development, manufacturing, and commercialization
Clinical trial optimization
Precision trial design, AI-assisted enrollment criteria, patient phenotyping and stratification, endpoint selection, and radiology report analysis.
Manufacturing and supply chain
Anomaly detection, predictive maintenance, visual quality control, stability data analysis, drug supply planning, shipment and inventory management, and CDMO monitoring.
Commercial operations
Dynamic forecasting models, launch planning, sales and operational planning, and sales analytics.
Synthetic control arms and digital twins
Tools designed to reduce the size and cost of clinical trials by supplementing placebo arms with synthetic controls built from existing datasets, and by simulating trial outcomes in virtual environments.
Alongside DrugTX.AI, our VR training systems build operational proficiency through immersive biomanufacturing training environments, interactive instrument simulation, and physics-based performance monitoring.
Ten thousand assets, scored against one framework.
Our Intelligent Asset Search and Prioritization system scans and evaluates over 10,000 global biopharmaceutical assets, drawing on published research, clinical trial outcomes, regulatory filings, and market intelligence, with indication-specific algorithms for rare disease, oncology, and other therapeutic areas.
It was through this system that we identified UNDC-005, by systematically analyzing the global cyclodextrin and lysosomal storage disease landscape. UNDC-005 scored highest on regulatory precedent, mechanistic differentiation, and capital efficiency. The prioritization was led by human pharmacodynamic data already in the record: 50% systemic absorption and statistically significant phospholipid clearance.
| Criterion | Priority | Rationale |
|---|---|---|
| Development phase | Highest | Later-stage assets carry lower risk and a clearer path to commercialization |
| Clinical indication and unmet medical need | High | Drives addressable market size and eligibility for expedited regulatory pathways such as Orphan Drug Designation and Fast Track |
| Novelty of mechanism of action | Medium | Signals first-in-class or best-in-class potential and competitive differentiation |
| Strength of supporting data and IP protection | Medium | Strong data indicates lower development risk; IP protection provides market exclusivity |
| Market potential | Medium | Directly impacts revenue generation and commercial attractiveness to partners |
| Clarity and maturity of asset description | Essential for evaluation | Indicates development readiness and simplifies due diligence |
Aligned with the FDA's draft guidance on AI
In January 2025 the FDA issued draft guidance titled Artificial Intelligence in the Development of Drug and Biological Products, establishing a risk-based, seven-step credibility assessment framework.
- Define the question of interest
- Define the context of use
- Assess model risk
- Develop a credibility assessment plan
- Execute the plan
- Document results
- Determine adequacy, with ongoing monitoring for data drift
Our platforms are designed to align with the principles in that guidance — context-of-use-specific model validation, fit-for-use data standards, and methodological transparency, so that AI-driven insights such as endpoint selection remain grounded in traceable biological evidence. We maintain protocols for ongoing monitoring to identify and mitigate model performance drift.