Learning protein sequence embeddings using information from structure

Tristan Bepler, Bonnie Berger

Abstract: We present a method for learning protein sequence embedding models using structural information in the form of global structural similarity between proteins and within protein residue-residue contacts.