Seattle Launches AI BioDesign to Propose New Proteins and Genes for Laboratory Testing

Seattle Launches AI BioDesign to Propose New Proteins and Genes for Laboratory Testing

On September 3, the Allen Institute, the University of Washington, and the Fred Hutch Cancer Center announced the AI BioDesign program. Its models will propose biological designs, and the results of successive laboratory experiments will guide each new round of work.

Over billions of years, evolution has produced only a fraction of all possible DNA sequences. In their joint announcement, the partners describe how they plan to explore other possibilities: a model proposes protein and gene sequences, then scientists create them and measure the results.

After each round, the measurements are returned to the model and used to select the next design. Parallel experiments should allow researchers to test many designs at once. The organizers plan to develop separate models for specific biological problems.

David Baker, one of the program’s scientific leads and director of the University of Washington’s Institute for Protein Design, describes the potential of synthetic biology, which involves creating and testing new biological systems, as follows:

“For the first time, the speed of AI is beginning to match the experimental capacity of synthetic biology. We can now ask what else is biologically possible and how we can test it, rather than limiting the question to what nature has already produced.”

The Allen Institute develops large, open scientific resources. The University of Washington works on protein design and DNA research, while the Fred Hutch Cancer Center studies cells and genomes.

The organizers plan to publish the models, datasets, laboratory tests, reagents, which are substances used in experiments, and benchmarks, which are standardized tasks for comparing methods, as open resources. Other research groups will be able to use them.