A major new collaboration between the Allen Institute, University of Washington, and Fred Hutch Cancer Center is taking AI drug discovery in a different direction: instead of just finding existing drugs faster, researchers will use AI to design completely new biomolecules that nature never created, then test them in living cells to validate the findings. The initiative, called AI BioDesign, represents a shift away from pure computational prediction toward what experts call "wet-lab validation," where AI-generated designs get tested in actual biological systems before moving forward.

What Makes AI BioDesign Different From Other AI Drug Discovery Tools?

Most AI tools in pharmaceutical development focus on speeding up the drug discovery pipeline itself, compressing timelines through target identification, molecule generation, and virtual screening. AI BioDesign takes a fundamentally different approach. Rather than hunting for new drugs, the program aims to understand the underlying principles of how biology works, then use that knowledge to create novel molecules that could serve as building blocks for future therapies.

The distinction matters because it addresses what researchers call the "data silo" problem. Pharmaceutical and biotech companies hoard proprietary data about their compound libraries and experimental results, keeping it locked away from the broader scientific community. Meanwhile, publicly available data comes from different labs using different methods and testing conditions, making it messy and inconsistent for training AI models. AI BioDesign sidesteps this bottleneck by generating new data through controlled experiments, then sharing the results openly to benefit the entire field.

"Our work, which will be shared openly, will advance how drug discovery workflows proceed," stated Dr. Justin English, director of strategy at the University of Washington Medicine Institute for Protein Design.

Dr. Justin English, Director of Strategy, UW Medicine Institute for Protein Design

How Will AI BioDesign Actually Create New Biomolecules?

The process unfolds in three key stages. First, AI algorithms will generate entirely new DNA sequences that code for biological molecules never seen in nature. Second, researchers will synthesize these molecules in the lab and run experiments to test their actual properties and functions. Third, the resulting data feeds back into the AI system, creating a feedback loop that continuously improves the models' understanding of how atoms and molecules interact.

This cycle addresses a critical weakness in current AI drug discovery: many tools make predictions based on patterns in historical data, but those predictions often fail when tested in real biology. By closing the loop between prediction and experimental validation, AI BioDesign aims to build models that actually understand the underlying rules of molecular behavior, not just statistical correlations.

Steps to Validate AI-Designed Molecules in Drug Discovery

• AI Sequence Generation: Machine learning algorithms analyze patterns in known biological sequences to design novel DNA and protein sequences that have never existed in nature.
• Experimental Synthesis and Testing: Researchers physically create the AI-designed molecules in the laboratory and measure their properties, binding affinity, and biological activity under controlled conditions.
• Data Integration and Model Refinement: Experimental results are fed back into the AI system to improve future predictions, creating a continuous learning cycle that strengthens the models' understanding of molecular principles.
• Functional Validation: Novel molecules undergo further analysis to confirm they perform their intended functions, such as binding to disease targets or triggering specific cellular responses.

The program is supported by the Fund for Science and Technology and brings together AI specialists and biologists across three major institutions. Dr. Sanjay Srivatsan, assistant professor at the Fred Hutch Cancer Center and principal investigator at AI BioDesign, emphasized that the initiative targets fundamental questions about how biology works at the atomic level.

"The goal of this program is to probe and define the principles around the logic that guides assembly and organization of living systems. How do proteins bind to DNA or a small molecule? What are the atomic features of a biomolecular surface that confer specificity, efficacy, or affinity? These are absolutely quintessential principles at the core of developing medicines," explained Dr. Sanjay Srivatsan.

Dr. Sanjay Srivatsan, Assistant Professor, Fred Hutch Cancer Center

Why Does Wet-Lab Validation Matter for AI-Driven Science?

The emphasis on experimental testing reflects a growing recognition that AI predictions alone are insufficient for drug development. As AI-enabled science becomes more prevalent, researchers increasingly need to validate computational findings against real biological systems before investing in expensive clinical trials. This is where companies like Revvity are positioning themselves strategically.

Revvity recently announced its acquisition of Human Cell Design, a France-based biotechnology company specializing in human cell models for metabolic disease research. The deal brings Revvity access to the EndoC-βH5 human pancreatic beta cell model, which researchers can use to test drug candidates and validate AI-derived predictions. The global cell analysis market is valued at approximately $37.3 billion in 2026 and is expected to grow at an annual rate of 11.7% through 2033, driven partly by the increasing need for sophisticated tools that bridge AI prediction and biological reality.

Revvity's integration of Human Cell Design's cell models with its own detection and screening technologies creates what the company calls "more integrated and human-relevant solutions." This combination allows researchers to generate and interpret complex cellular data, turning AI predictions into validated biological insights.

What Are the Broader Implications for Drug Manufacturing?

Beyond drug discovery, AI BioDesign has implications for pharmaceutical manufacturing. Dr. English noted that the program will explore how to create entirely new biological systems that could overcome current limitations in chemical synthesis and production speed. This matters because discovering a promising drug candidate is only half the battle; scaling up production to reach patients requires reliable, efficient manufacturing processes.

"Our program will also address fundamental questions around how to create entirely new biological systems that can unlock and overcome limits that currently exist in biological processes, from chemical synthesis by enzymes to cellular division speed," added Dr. Justin English.

Dr. Justin English, Director of Strategy, UW Medicine Institute for Protein Design

The five-year collaboration represents a recognition that AI's role in drug discovery is evolving. Rather than replacing traditional research, AI is becoming a tool for understanding nature's design principles and applying them to create solutions that nature never produced. By combining computational power with rigorous experimental validation and open data sharing, AI BioDesign aims to accelerate the entire pipeline from molecular design to therapeutic development, while building a foundation of shared knowledge that benefits the broader scientific community.