FoldToken: Learning Protein Language via Vector Quantization and Beyond

TL;DR

FoldToken introduces a novel discrete representation for protein sequences and structures, enabling a unified protein language model that improves tasks like backbone inpainting and antibody design.

Contribution

The paper proposes FoldTokenizer and SoftCVQ to discretize protein data, creating a new language model for sequence-structure co-generation.

Findings

Effective protein sequence-structure discretization

Successful application to backbone inpainting and antibody design

First GPT-style model for protein sequence-structure generation

Abstract

Is there a foreign language describing protein sequences and structures simultaneously? Protein structures, represented by continuous 3D points, have long posed a challenge due to the contrasting modeling paradigms of discrete sequences. We introduce \textbf{FoldTokenizer} to represent protein sequence-structure as discrete symbols. This innovative approach involves projecting residue types and structures into a discrete space, guided by a reconstruction loss for information preservation. We refer to the learned discrete symbols as \textbf{FoldToken}, and the sequence of FoldTokens serves as a new protein language, transforming the protein sequence-structure into a unified modality. We apply the created protein language on general backbone inpainting and antibody design tasks, building the first GPT-style model (\textbf{FoldGPT}) for sequence-structure co-generation with promising results. Key to our success is the substantial enhancement of the vector quantization module, Soft Conditional Vector Quantization (\textbf{SoftCVQ}).