Verge Labs Launches Disease Biology AI World Model to Accelerate Drug Trials
Draper Associates announced the official launch of Verge Labs, aimed at addressing the 90% failure rate of drug trials, often due to incorrect target and patient matching.
The company has spent a decade building the extensive VergeDB database and has trained a foundational model on this basis to generate "virtual biopsies," simulating the molecular landscape of patients' brains through simple blood samples to predict therapy effects, disease progression, and biomarker responses.
Event-driven AI empowers precision medicine acceleration, with funding and talent flowing into disease modeling platforms, benefiting neuroscience drug development through pre-trial screening, putting pressure on traditional high-failure-rate trial models.
Source: Public Information
ABAB AI Insight
Verge Labs, led by Alice Xinli Zhang, has long accumulated multi-omics and clinical data to build VergeDB, now applying the foundational model to a human disease biology "world model." The first AI-discovered asset has completed Phase 1b trials and will feed data back into the system, while also bringing in top talents from Altos Labs, Calico, and others.
On the capital path, the company reduces clinical trial risks through data flywheels and virtual predictions, motivated to replicate the precision medicine path of oncology in neuroscience, focusing resources on blood-driven molecular simulations rather than large-scale blind trials.
Similar to early genomics companies or Flatiron Health's data-driven tumor decision-making path, it is currently transitioning neuroscience diseases from empirical trials to predictive modeling.
Essentially, this represents a technological substitution: virtual biopsies and world models replace parts of traditional target validation processes, as precise matching before trials can significantly reduce failure rates, thereby reconstructing the cost and success rate structure of drug development.
ABAB News · Cognitive Laws
Incorrect patients waste trials more than incorrect targets.
Data flywheels are the true moat of AI pharmaceuticals.
Virtual biopsies turn blood samples into brain maps.