Rapid identification of improved antibody variants; plus peptide-assisted genome editing and more

Stanford University scientists presented in Nature Biotechnology a model to increase the target-binding affinity and therapeutic potential of antibodies and other proteins by suggesting patterns that are likely to occur in natural proteins without providing the model with any information about the target antigen, binding specificity or protein structure.

The authors used the protein language model to improve the affinity of seven antibodies that bind to antigens from coronavirus, ebolavirus and influenza A virus after measuring 20 or fewer new variants of each antibody across two rounds of evolution, which the authors said “represents unprecedented efficiency for machine-learning-guided evolution.”...