Summary
Protein language model attention maps can be used to predict allosteric networks in proteins more effectively than coevolutionary analysis via Potts models ((1), (2)). Note that only the latter authors found that specific attention heads were useful for predicting allostery.
Figures
Ref (1)
1.
Kannan GR, Hie BL, Kim PS. Single-Sequence, Structure Free Allosteric Residue Prediction with Protein Language Models. openRxiv; 2024. Available from: https://doi.org/10.1101/2024.10.03.616547
2.
Dong T, Kan C, Devkota K, Singh R. Allo-Allo: Data-efficient prediction of allosteric sites. openRxiv; 2024. Available from: https://doi.org/10.1101/2024.09.28.615583