This reframes the folding problem as: what determines the burial of the hard-to-predict core residues? The core identity score is available on GitHub with a Google Colab notebook. Try it on your own structures! 8/8 Link: github.com/agrigas115/core_identity_score
Predict a protein structure's backbone accuracy (average LDDT) by comparing its core residues to the core predicted using ESM2 embeddings - agrigas115/core_identity_score
