Atom-by-atom protein generation and beyond with language models

TL;DR

This paper demonstrates that chemical language models can generate proteins atom-by-atom, enabling the design of novel proteins and biomolecules beyond the constraints of the genetic code, including unnatural amino acids and protein-drug conjugates.

Contribution

It introduces a method for atom-level protein generation using language models, bridging chemical and biological representations for advanced biomolecular design.

Findings

Language models can generate proteins with atom-level detail.

Models can create proteins with unnatural amino acids.

Simultaneous exploration of chemical and protein space is possible.

Abstract

Protein language models learn powerful representations directly from sequences of amino acids. However, they are constrained to generate proteins with only the set of amino acids represented in their vocabulary. In contrast, chemical language models learn atom-level representations of smaller molecules that include every atom, bond, and ring. In this work, we show that chemical language models can learn atom-level representations of proteins enabling protein generation unconstrained to the standard genetic code and far beyond it. In doing so, we show that language models can generate entire proteins atom by atom -- effectively learning the multiple hierarchical layers of molecular information that define proteins from their primary sequence to their secondary, and tertiary structure. We demonstrate language models are able to explore beyond protein space -- generating proteins with modified sidechains that form unnatural amino acids. Even further, we find that language models can explore chemical space and protein space simultaneously and generate novel examples of protein-drug conjugates. The results demonstrate the potential for biomolecular design at the atom level using language models.