P(all-atom) Is Unlocking New Path For Protein Design

Wei Qu, Jiawei Guan, Rui Ma, Ke Zhai, Weikun Wu, Haobo Wang
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:50786-50816, 2025.

Abstract

We introduce Pallatom, an innovative protein generation model capable of producing protein structures with all-atom coordinates. Pallatom directly learns and models the joint distribution $P(\textit{structure}, \textit{seq})$ by focusing on $P(\textit{all-atom})$, effectively addressing the interdependence between sequence and structure in protein generation. To achieve this, we propose a novel network architecture specifically designed for all-atom protein generation. Our model employs a dual-track framework that tokenizes proteins into token-level and atomic-level representations, integrating them through a multi-layer decoding process with "traversing" representations and recycling mechanism. We also introduce the $\texttt{atom14}$ representation method, which unifies the description of unknown side-chain coordinates, ensuring high fidelity between the generated all-atom conformation and its physical structure. Experimental results demonstrate that Pallatom excels in key metrics of protein design, including designability, diversity, and novelty, showing significant improvements across the board. Our model not only enhances the accuracy of protein generation but also exhibits excellent sampling efficiency, paving the way for future applications in larger and more complex systems.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-qu25c, title = {P(all-atom) Is Unlocking New Path For Protein Design}, author = {Qu, Wei and Guan, Jiawei and Ma, Rui and Zhai, Ke and Wu, Weikun and Wang, Haobo}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {50786--50816}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/qu25c/qu25c.pdf}, url = {https://proceedings.mlr.press/v267/qu25c.html}, abstract = {We introduce Pallatom, an innovative protein generation model capable of producing protein structures with all-atom coordinates. Pallatom directly learns and models the joint distribution $P(\textit{structure}, \textit{seq})$ by focusing on $P(\textit{all-atom})$, effectively addressing the interdependence between sequence and structure in protein generation. To achieve this, we propose a novel network architecture specifically designed for all-atom protein generation. Our model employs a dual-track framework that tokenizes proteins into token-level and atomic-level representations, integrating them through a multi-layer decoding process with "traversing" representations and recycling mechanism. We also introduce the $\texttt{atom14}$ representation method, which unifies the description of unknown side-chain coordinates, ensuring high fidelity between the generated all-atom conformation and its physical structure. Experimental results demonstrate that Pallatom excels in key metrics of protein design, including designability, diversity, and novelty, showing significant improvements across the board. Our model not only enhances the accuracy of protein generation but also exhibits excellent sampling efficiency, paving the way for future applications in larger and more complex systems.} }
Endnote
%0 Conference Paper %T P(all-atom) Is Unlocking New Path For Protein Design %A Wei Qu %A Jiawei Guan %A Rui Ma %A Ke Zhai %A Weikun Wu %A Haobo Wang %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-qu25c %I PMLR %P 50786--50816 %U https://proceedings.mlr.press/v267/qu25c.html %V 267 %X We introduce Pallatom, an innovative protein generation model capable of producing protein structures with all-atom coordinates. Pallatom directly learns and models the joint distribution $P(\textit{structure}, \textit{seq})$ by focusing on $P(\textit{all-atom})$, effectively addressing the interdependence between sequence and structure in protein generation. To achieve this, we propose a novel network architecture specifically designed for all-atom protein generation. Our model employs a dual-track framework that tokenizes proteins into token-level and atomic-level representations, integrating them through a multi-layer decoding process with "traversing" representations and recycling mechanism. We also introduce the $\texttt{atom14}$ representation method, which unifies the description of unknown side-chain coordinates, ensuring high fidelity between the generated all-atom conformation and its physical structure. Experimental results demonstrate that Pallatom excels in key metrics of protein design, including designability, diversity, and novelty, showing significant improvements across the board. Our model not only enhances the accuracy of protein generation but also exhibits excellent sampling efficiency, paving the way for future applications in larger and more complex systems.
APA
Qu, W., Guan, J., Ma, R., Zhai, K., Wu, W. & Wang, H.. (2025). P(all-atom) Is Unlocking New Path For Protein Design. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:50786-50816 Available from https://proceedings.mlr.press/v267/qu25c.html.

Related Material