Summary
Deep learning models for protein design using natural language as input are more effective when predicting protein structure rather than sequence (1). Their model uses either amino acids or (SaProt) tokens to represent structure.
Figures
Ref (1)
1.
Dai F, You S, Zhu Y, Gao Y, Fu L, Zhou X, et al. Toward De Novo Protein Design from Natural Language. openRxiv; 2024. Available from: https://doi.org/10.1101/2024.08.01.606258