For decades, the pharmaceutical industry has operated under a rigid, linear philosophy: discover a potent molecule first, and figure out if it is "drug-like" later. In this traditional model, comprehensive Absorption, Distribution, Metabolism, and Excretion (ADME) profiling—the crucial data that determines how a drug behaves in a human body—was relegated to the "lead optimization" stage. By then, companies had already invested months of labor and significant capital into a chemical series, only to discover, often too late, that the lead compounds were metabolically unstable, poorly absorbed, or required toxic doses.
A new strategic partnership between Ginkgo Datapoints, Tangible Scientific, and Inductive Bio aims to dismantle this inefficient, high-risk pipeline. By launching "ADME-One," a unified, high-throughput platform, these three organizations are attempting to pull pharmacokinetic (PK) projections out of the end-stage bottleneck and into the earliest moments of hit identification.
The Core Concept: Front-Loading Pharmacokinetics
The fundamental value proposition of ADME-One is simplicity through integration. The platform combines Ginkgo’s automated Tier 1 assay suite, Tangible Scientific’s sophisticated compound management, and Inductive Bio’s machine-learning-driven human PK projection software.
In the conventional drug discovery cycle, medicinal chemists are often flying blind, focusing heavily on potency—the ability of a molecule to bind to its target—while lacking visibility into whether that same molecule will survive the gauntlet of the human liver or cross biological barriers. ADME-One flips this script. By providing a comprehensive read on potency, ADME parameters, and projected human PK before a program is fully committed to a specific chemical scaffold, the platform allows chemists to "fail fast" or iterate intelligently.
Chronology of the Development
The genesis of ADME-One emerged from a shared frustration among industry experts regarding the widening gap between computational prediction and empirical reality.
- The Recognition of the Gap: Leaders at Inductive Bio identified that while AI models were becoming more adept at predicting binding affinity, the "downstream" reality of ADME remained a manual, slow-moving, and fragmented process.
- Building the Consortium: The partners spent months performing "behind-the-scenes" validation to ensure that high-throughput automation did not sacrifice data quality. This involved calibrating Ginkgo’s robotic laboratories to meet the rigorous standards of medicinal chemistry teams accustomed to bespoke, small-scale experiments.
- Logistical Integration: Tangible Scientific provided the critical link by managing the physical custody of compounds. By handling the intake, plating, and real-time tracking of orders, they removed the "logistical friction" that often causes delays between design, synthesis, and testing.
- Market Launch: The platform was officially introduced to offer a domestic, high-speed alternative to offshore contract research organizations (CROs), specifically timed to address current geopolitical concerns regarding supply chain security and the BIOSECURE Act.
Supporting Data: Why "Dose" is the North Star
A critical driver for the ADME-One platform is the industry’s increasing focus on human dose optimization. As Dr. Alex Taylor, head of medicinal chemistry at Inductive Bio, emphasizes, "Experienced medicinal chemists will tell you up front that dose is ultimately the thing you want to optimize for."
The clinical rationale is supported by decades of pharmacological literature. Studies published in journals like Hepatology have consistently highlighted the "rule-of-two," where high daily doses—particularly those coupled with high lipophilicity—are strongly correlated with an elevated risk of drug-induced liver injury (DILI). Registry studies have further reinforced that compounds requiring doses exceeding 50 mg per day carry significantly higher risks of severe adverse events compared to those requiring less than 10 mg.
However, the industry has historically struggled to integrate these insights into the discovery process because the data required to calculate a "human dose projection" was rarely available in the hit-identification phase. By automating the five core Tier 1 assays—microsomal stability, cell permeability, kinetic solubility, CYP inhibition, and plasma protein binding—ADME-One provides the necessary variables to move from abstract potency to concrete dosing estimates.
Official Perspectives and Technical Strategy
The technical brilliance of the platform lies in its ability to reconcile different chemical profiles. Dr. Taylor points to the historical example of triazole antifungals: Fluconazole and Itraconazole represent two ends of the spectrum in terms of lipophilicity, protein binding, and clearance pathways. Despite their radically different profiles, both became clinical successes.
"Sometimes, compounds you think aren’t good enough to go forward… actually have a balance of all the properties such that they could go forward," Taylor notes. ADME-One helps identify this "balance" early on, preventing the premature discarding of valuable chemical matter that might otherwise be dismissed by rigid, single-parameter screening.
Addressing the Data Security Dilemma
A recurring concern for any consortium-based AI model is intellectual property protection. How can Inductive Bio improve its global models using client data without leaking proprietary chemical structures?
The company has implemented a dual-layered architecture. The "consortium" model acts as a legal and technical firewall. When a client contributes their experimental data, Inductive Bio uses it to fine-tune a "local" model specific to that client, which sits atop a robust "global" model. Crucially, the engineering team has designed the system to be immune to "reverse-engineering"—meaning no single partner can inspect the consortium’s aggregate data to deduce the chemistry of a competitor.
Implications for the Future of Drug Discovery
The arrival of ADME-One signals several major shifts in the pharmaceutical landscape:
1. The Rise of "Onshore" Resilience
With the enactment of the BIOSECURE Act and broader concerns about data sovereignty, there is a clear trend toward bringing preclinical research back to the United States and Europe. ADME-One’s commitment to a domestic workflow, combined with a turnaround time of days rather than weeks, positions it as a competitive threat to traditional offshore CRO models that have dominated the market for years.
2. The "Virtuous Cycle" of Machine Learning
The platform creates a self-reinforcing feedback loop. As more clients use the platform, the global model gains exposure to a wider variety of chemical matter. This, in turn, makes the predictive power of the model stronger for every subsequent user. This "flywheel" effect is expected to accelerate as the platform scales, potentially shortening the duration of the hit-to-lead phase across the industry.
3. Redefining the Role of AI
Perhaps the most grounded insight from the developers of ADME-One is the realistic role assigned to artificial intelligence. Dr. Taylor is careful to avoid "AI-hype," stating, "Drug discovery is science at the end of the day, and science is not engineering. There’s still so much that needs to happen empirically."
By positioning AI as a prioritization tool rather than a replacement for experimentation, the platform aligns itself with the practical realities of the lab. It doesn’t eliminate the need for synthesis or testing; it simply ensures that the compounds being synthesized and tested are the ones most likely to succeed.
4. Economic Discipline
In an era where venture capital is increasingly selective, the "lean" mantra has become a survival imperative. The cost-effectiveness of ADME-One, driven by robotic automation and the ability to process entire plates of compounds in parallel, allows small-to-mid-sized biotech firms to operate with the same data-rich efficiency as Big Pharma.
Conclusion
The launch of ADME-One is more than a technical upgrade; it is a fundamental shift in the economics and logistics of drug discovery. By moving the "decision point" for pharmacokinetic success earlier in the pipeline, Ginkgo, Tangible, and Inductive Bio are offering a path to reduce the astronomical costs of failure in clinical development.
As the industry moves toward a future where "data-first" discovery is the norm, the ability to rapidly cycle between computational prediction and experimental validation will likely become the primary differentiator between firms that stagnate and those that consistently deliver breakthrough therapies. The "virtuous cycle" promised by the ADME-One consortium suggests that the next generation of drugs may not just be discovered faster—they may be safer, more effective, and more precisely dosed than ever before.
About the Author
