GSK tests Chai-designed protein binders
GSK found protein binders for every target tested among candidates proposed by Chai Discovery models, then expanded the work into a collaboration.
GSK tested proteins proposed by Chai Discovery models
GSK tested proteins proposed by Chai Discovery models in its own laboratory and found proteins that bind to each of several targets tested. The companies announced their collaboration on October 9, after this evaluation.
Chai Discovery develops artificial intelligence models for protein research. The search for a protein drug begins with a target: a molecule in the body whose activity researchers want to influence. The model proposes protein sequences that could bind to that molecule. A laboratory then checks whether binding occurs and measures its strength. Chai proposed candidates for GSK without training the model separately for each target.
GSK therefore obtained experimental evidence that proposed proteins bind to the selected targets. The source material gives no exact target count, candidate success rate, or numerical binding measurements. Finding a binder for each target does not mean that every proposed candidate worked. Binding alone also cannot establish safety, therapeutic efficacy, or readiness for human trials.
Following this evaluation, the companies expanded their work together. GSK will gain access to Chai models to predict how proteins fold and what shape they take, and to design new protein candidates for its drug discovery programs.
Open the related Eternal Search page
Sources
[1] businesswire.com
[2] contractpharma.com
This item gives readers a concrete reason to inspect Chai Discovery in Eternal Search: a future partner tested its proposed proteins in its own laboratory, while the missing target counts, success rates, and binding measurements keep the result in proper scope.