Measured acceleration · Biology · Medicine
Protein language models guide antibody affinity maturation
Antibody affinity improvements found with small language-model-guided variant screens.
Summary
The Nature Biotechnology paper reports language-model-guided affinity maturation of seven antibodies. By screening 20 or fewer variants per antibody across two rounds of laboratory evolution, the authors improved binding affinities for several clinically relevant and unmatured antibodies and tested properties such as thermostability and pseudovirus neutralization.
AI role
Guided antibody variant selection for laboratory evolution across clinically relevant and unmatured antibodies.
Narrative role
This is a concrete process-acceleration event because it links AI guidance to reduced search effort in therapeutic protein engineering.
Caveat
The workflow is demonstrated on selected antibodies and does not remove broader development and manufacturability constraints.