In the high-stakes arena of drug discovery, the timeline from a promising hit to a viable clinical candidate is often a graveyard of failed projects. For decades, the industry has relied on a "serial" approach: identify a hit, optimize it for potency, and only then—often after months of expensive synthesis cycles—invest in comprehensive Absorption, Distribution, Metabolism, and Excretion (ADME) profiling. This traditional workflow frequently results in a "late-stage surprise," where molecules that perform well in a test tube fail spectacularly once human pharmacokinetic (PK) projections are finally calculated.
A new strategic partnership between Ginkgo Datapoints, Tangible Scientific, and Inductive Bio aims to dismantle this paradigm. By launching "ADME-One," a high-throughput, integrated service platform, the trio is betting that AI-driven automation can shift the burden of pharmacokinetic analysis from the late-stage lead optimization phase to the very beginning of the hit identification process.
The Paradigm Shift: From Late-Stage Gatekeeping to Early-Stage Insight
The fundamental premise of ADME-One is to make sophisticated, predictive pharmacology accessible during the earliest stages of chemical exploration. Historically, medicinal chemists were forced to prioritize potency above all else during hit ID, essentially "flying blind" regarding how those molecules would behave in a biological system.
"We asked: 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.
By integrating Ginkgo’s automated Tier 1 assays—including microsomal stability, cell permeability, kinetic solubility, CYP inhibition, and plasma protein binding—with Tangible’s high-precision compound logistics and Inductive’s proprietary predictive modeling, the consortium has created a "one-stop-shop" service. The goal is to provide chemists with a holistic view of a molecule’s viability long before a program becomes overly committed to a specific, potentially flawed, chemical series.
A Chronology of Collaboration
The development of ADME-One was not an overnight success but the result of significant behind-the-scenes validation.
- Foundation Phase: Inductive Bio established its "Compass" platform, designed to roll disparate experimental readouts into a unified human PK projection.
- Logistics Integration: Tangible Scientific refined its workflow to handle the physical custody of compounds, optimizing intake, plating, and real-time tracking to ensure that the "chain of custody" for small-molecule discovery remained seamless.
- Technical Validation: Ginkgo Datapoints utilized its automated laboratory infrastructure in Boston to standardize the Tier 1 assays. The partners spent months performing cross-validation to ensure that the high-throughput, automated results matched the gold-standard data quality expected in traditional research settings.
- Launch: The formal launch of the ADME-One platform marks the culmination of this tripartite effort, offering a service that promises to return high-fidelity data in days, rather than the weeks typically required by conventional outsourcing models.
Supporting Data: Why "Dose" is the North Star
The urgency of early-stage ADME profiling is underscored by a growing recognition in medicinal chemistry that the human dose is the ultimate metric of success. Experienced practitioners know that a drug’s physical properties, such as lipophilicity and molecular weight, are inextricably linked to its safety profile.
Research from FDA-affiliated scientists has long highlighted the "rule-of-two," where high daily doses, particularly in combination with high lipophilicity, correlate with an increased risk of drug-induced liver injury (DILI). Furthermore, registry studies have indicated that compounds requiring doses of 50 mg per day or higher face significantly elevated risks of clinical failure, including liver-related adverse events, compared to those with lower dosing requirements.
"Experienced medicinal chemists will tell you up front that dose is ultimately the thing you want to optimize for," Taylor notes. However, the data required to estimate a human dose—potency combined with ADME parameters—has historically been too expensive and time-consuming to gather at scale.
The ADME-One platform seeks to challenge this by providing a "virtuous cycle" of data. By testing more compounds early, teams can avoid the pitfalls seen with past drugs. For instance, the starkly different profiles of fluconazole (small, polar, minimal protein binding) and itraconazole (highly lipophilic, extensively protein-bound, hepatic metabolism) demonstrate that there is no "perfect" profile for a drug. Success is defined by the balance of these properties. ADME-One provides the granular data necessary for chemists to understand that balance from the start.
The Economics of Onshore Efficiency
The drive toward ADME-One is as much about economic survival as it is about scientific rigor. The pharmaceutical and biotechnology sectors are currently under intense pressure to remain "lean" and "cost-conscious." This financial reality has historically forced teams to rely on a "screening-funnel" approach, where only a tiny fraction of molecules receive deep profiling.
Furthermore, geopolitical and regulatory shifts have spurred a return to domestic data generation. The BIOSECURE Act and an increasing demand for "data sovereignty"—keeping sensitive research data within the U.S. and European borders—have made offshore contract research organizations (CROs) less attractive for high-value early discovery. ADME-One positions itself as a domestic, cost-competitive alternative that delivers speed without compromising the security or quality of the data.
Data Security and the Consortium Model
Perhaps the most complex hurdle for any AI-driven drug discovery partnership is the "data silo" problem. How can companies train powerful machine learning (ML) models on proprietary chemistry without risking the intellectual property of their clients?
Inductive Bio manages this through a separate, highly secure consortium model. "The founders put together a consortium model, a legal framework where all our partners can pool their data securely, and no partner can see anyone else’s data," Taylor explains.
This infrastructure is designed to prevent "reverse-engineering." The system uses a global model trained on the aggregate, anonymized data pool, which is then fine-tuned with a "local model" specific to the client’s own data. This approach offers the best of both worlds: the predictive power of a vast, cross-industry dataset and the specialized performance of a model tailored to a specific project’s unique chemical matter.
Implications for the Future of Drug Discovery
The arrival of ADME-One suggests that the future of drug discovery lies in the integration of three distinct pillars: high-quality experimental data, sophisticated predictive modeling, and efficient logistics.
1. The Death of the "Slow-Fail"
By surfacing ADME and PK issues during the hit identification stage, teams can terminate non-viable programs months earlier. This "fail fast, fail cheap" mentality is the hallmark of modern, efficient R&D.
2. The Rise of the "Virtuous Cycle"
As more programs utilize the ADME-One platform, the global ML models will continue to gain breadth in chemical space. This creates a flywheel effect: better models lead to better initial compound designs, which lead to more accurate empirical results, which in turn train the models to be even more precise for the next project.
3. Science vs. Engineering
Despite the excitement surrounding AI, the partners remain grounded in the reality of the lab. As Taylor cautions, "Drug discovery is science at the end of the day, and science is not engineering." The platform does not replace the chemist; it serves as a prioritization tool. It helps teams decide which compounds are worth the significant time and expense of synthesis.
4. A New Standard for CROs
The emergence of this consortium model sets a new bar for service providers. Future CROs will likely be judged not just on the accuracy of their individual assays, but on their ability to integrate into the digital workflows of their clients, offering actionable insights rather than just raw data.
Conclusion
The collaboration between Ginkgo, Tangible, and Inductive Bio represents a significant evolution in the drug discovery ecosystem. By democratizing access to high-throughput, integrated pharmacokinetic projections, the ADME-One platform challenges the industry to rethink its traditional, linear workflows.
For the medicinal chemist, this means fewer wasted synthesis cycles and a clearer path to identifying compounds that not only bind to a target but also possess the physical properties necessary to succeed in a human body. As the pharmaceutical industry continues to grapple with the rising costs of R&D and the persistent challenge of high attrition rates, solutions that prioritize early-stage intelligence may well become the standard for the next generation of drug discovery.
About the Author
