In the high-stakes world of pharmaceutical development, the traditional "fail fast" mantra has long been hindered by a bottleneck: the delay of comprehensive Absorption, Distribution, Metabolism, and Excretion (ADME) profiling until the lead optimization phase. By the time many small-molecule programs reach this stage, companies have already committed significant time and capital to chemical series that may harbor fundamental pharmacokinetic flaws.
A new collaborative venture—comprising Ginkgo Datapoints, Tangible Scientific, and Inductive Bio—is seeking to dismantle this paradigm. By launching "ADME-One," a high-throughput, AI-driven platform, the consortium aims to move critical ADME and human pharmacokinetic (PK) projections from the end of the discovery funnel to the very beginning: the hit identification phase.
The Core Philosophy: Shifting the Paradigm
The conventional approach to drug discovery has historically treated ADME profiling as a luxury of the "lead optimization" phase, performed only after a series has shown initial potency. This strategy, however, often leads to the "sunk cost" fallacy, where teams continue to pour resources into compounds that are ultimately doomed by poor metabolic stability or unfavorable dosing profiles.
The ADME-One platform seeks to flip this model. By integrating automated laboratory assays, advanced compound management, and machine learning-driven pharmacokinetic forecasting, the partners aim to provide medicinal chemists with a holistic view of a compound’s potential before months of synthesis cycles have passed.
"The guiding question was: Could we pull together all the assays needed to get your first projection of human PK at a price point where you’d now be doing this on most, if not all, of the compounds coming through?" explains Alex Taylor, Ph.D., Head of Medicinal Chemistry at Inductive Bio.
A Triple-Threat Collaboration
The strength of ADME-One lies in the specialized contributions of its three founding partners, each addressing a specific pain point in the preclinical supply chain:
- Ginkgo Datapoints: Providing the infrastructure for execution, Ginkgo runs five Tier 1 assays—microsomal stability, cell permeability, kinetic solubility, CYP inhibition, and plasma protein binding—within their fully automated Boston-based laboratory.
- Tangible Scientific: Acting as the logistical backbone, Tangible Scientific manages the physical movement of compounds. They handle intake, plate preparation, and real-time tracking, ensuring that the integrity of the samples remains uncompromised throughout the high-throughput process.
- Inductive Bio: The intellectual engine of the partnership, Inductive Bio utilizes its proprietary "Compass" platform to synthesize the raw experimental data into actionable human PK projections. This allows teams to rank compounds not just by their raw potency, but by their likelihood of becoming a viable, well-tolerated drug.
Chronology of the Development and Launch
The genesis of the ADME-One platform reflects a broader trend in the life sciences: the convergence of robotics, data science, and chemical biology.
- The Conceptual Phase: Recognizing that the industry was facing mounting pressure to reduce costs while improving the quality of clinical candidates, the teams began identifying the gaps in current CRO-reliant models. The focus was on "data sovereignty" and "onshore efficiency."
- Validation and Integration: Before the public launch, the partners engaged in extensive "behind-the-scenes" validation. According to Dr. Taylor, the primary challenge was ensuring that the transition to a high-throughput, automated environment did not compromise the gold-standard data quality that medicinal chemists require.
- The Launch: The formal announcement marks the transition of this technology from an internal tool to a service-based platform, designed to compete with—and potentially outperform—offshore CRO models in terms of speed, cost, and reliability.
Supporting Data and the "Dose-First" Mindset
The shift toward early-stage PK modeling is not merely a technical preference; it is a clinical imperative. Medicinal chemistry literature has long suggested that "dose" is the ultimate objective of optimization, yet it is often the hardest parameter to predict early.
The Risks of High-Dose Drugs
Data from the FDA and various registry studies underscore the risks associated with high daily doses. Research published in Hepatology has highlighted the "rule-of-two," noting that high daily doses coupled with high lipophilicity are correlated with an elevated risk of drug-induced liver injury (DILI). Furthermore, studies indicate that drugs requiring doses of 50 mg or higher per day are statistically more likely to result in adverse events, including liver failure, compared to their low-dose counterparts.
By integrating ADME and potency data early, the ADME-One platform allows chemists to steer their designs away from high-dose, high-risk profiles before they move into expensive clinical trials.
Balancing Chemical Complexity
The platform also addresses the common pitfall of discarding "imperfect" compounds too early. Dr. Taylor points to the evolution of triazole antifungals as a case study. Molecules like fluconazole and itraconazole possess drastically different ADME profiles—one polar and renally cleared, the other highly lipophilic and hepatically metabolized. Yet, both proved successful. ADME-One provides the nuance necessary to see that a compound’s "balance" of properties is more important than its score in any single isolated assay.
Implications for the Industry
The introduction of ADME-One carries significant implications for both small-molecule startups and established pharma companies facing increased scrutiny under the BIOSECURE Act and the general push for more robust supply chains.
The Financial Argument
In an era where "staying lean" is the primary mandate, the ability to obtain high-quality, actionable data at a price point lower than traditional offshore CROs is a disruptive value proposition. By utilizing robotic automation, the partnership has effectively turned the process of running assays into a commodity, allowing firms to test more compounds for less money.
The Consortium Model and Data Security
A critical concern for any company sharing data with an AI platform is the risk of intellectual property leakage. Inductive Bio has addressed this through a sophisticated "consortium model."
The data architecture is designed such that participating partners pool their anonymized data to train global machine learning models, which in turn improves the predictive power of the system for all users. However, each user’s specific chemistry is protected by a legal and technical firewall. When a customer adds their proprietary data, the system fine-tunes a "local" model on top of the global one. This ensures that a client gains the benefit of the industry-wide "flywheel" of data without ever exposing their confidential structures to other entities.
Future Outlook: A Virtuous Cycle
The ultimate goal for the ADME-One partners is to foster a "virtuous cycle" in drug discovery. As more compounds are synthesized and tested through the platform, the global models become more accurate. This, in turn, informs better design choices in the next round of synthesis.
"Drug discovery is science at the end of the day, and science is not engineering," notes Dr. Taylor. "There’s still so much that needs to happen empirically."
The team remains grounded in the reality that AI cannot replace the chemist; it can only act as a compass. By narrowing the field of candidates to those most likely to succeed in a human body, the platform promises to make the drug discovery process faster, more cost-effective, and—most importantly—more likely to produce safer, more effective medicines.
As the industry continues to navigate the complexities of modern pharmacology, the shift toward integrating ADME and PK projections into the hit-ID phase appears not just as a competitive advantage, but as a necessary evolution of the craft. Whether the ADME-One platform will become the new industry standard remains to be seen, but the intent behind the consortium—to bring better, more data-driven decision-making to the earliest stages of research—is a welcome step toward a more efficient future for drug development.
About the Author
