Closed Benstime closed 2 years ago
Correct, the model does not take amino acid sequences as input. The model is for creating (or ``hallucinating'', if you will) a new backbone structure that specifies a protein shape, starting from noise. This shape can be then engineered using tools such as ESM-IF1.
so what can the model be used for?