Summary

Protein language model embeddings are more predictive of homology than catalytic efficiency (1). In fact, ESM embeddings were slightly better than random and worse than one-hot encodings at identifying highly active adenylate kinase variants.

Figures

Ref (1)

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