Summary
Closely-related enzymes can have disparate catalytic efficiencies, and their values correlate only weakly with sequence similarity (1). For example, adenylate kinase values span three orders of magnitude. This makes it difficult for protein language models to effectively predict catalytic properties in a zero-shot setting.
Figures
Ref (1); vertical line of top figure refers to the consensus sequence
1.
Muir DF, Asper GPR, Notin P, Posner JA, Marks DS, Keiser MJ, et al. Evolutionary-scale enzymology enables exploration of a rugged catalytic landscape. Science. 2025;388(6752). Available from: https://doi.org/10.1126/science.adu1058