Jointly Embedding Protein Structures and Sequences through Residue Level Alignment

Presented on August 14th 2024 by Foster Birnbaum
abstract: We propose Residue Level Alignment (RLA) - a self-supervised objective for aligning structure and sequence embedding spaces. Our framework improves each individual encoder and enables us to measure the similarity between a structure and sequence by comparing their RLA embeddings. We show that RLA similarity scores can select true protein binders from sets of designed binders and that RLA is more versatile and significantly faster than existing design scoring methods.