P(all-atom) Is Unlocking New Path For Protein Design
Wei Qu, Jiawei Guan, Rui Ma, Ke Zhai, Weikun Wu, Haobo Wang
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 (structure, seq) by focusing on P (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 residuelevel and atomic-level representations, integrating them through a multi-layer decoding process with "traversing" representations and recycling mechanism. We also introduce the 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.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext 2360c6e2-ee9a-4fbc-b2d1-e9e0e6945415Cited by top-tier papers9
- La-Proteina: Atomistic Protein Generation via Partially Latent Flow MatchingTomas Geffner, Kieran Didi, Zhonglin Cao, Danny Reidenbach et al.ICLR 2026 · 57 citations
- SimpleFold: Folding Proteins is Simpler than You ThinkYuyang Wang, Jiarui Lu, Navdeep Jaitly, Joshua M. Susskind et al.ICLR 2026 · 29 citations
- Greater than the Sum of Its Parts: Building Substructure into Protein Encoding ModelsRobert Calef, Arthur Liang, Manolis Kellis, Marinka ZitnikICLR 2026 · 2 citations
- Protein Autoregressive Modeling via Multiscale Structure GenerationYanru Qu, Cheng-Yen Hsieh, Zaixiang Zheng, Ge Liu et al.ICML 2026
- Proteina: Scaling Flow-based Protein Structure Generative ModelsTomas Geffner, Kieran Didi, Zuobai Zhang, Danny Reidenbach et al.ICLR 2025
Builds on9
- E(n) Equivariant Graph Neural NetworksVictor Garcia Satorras, Emiel Hoogeboom, Max WellingICML 2021 · 1,432 citations
- Learning inverse folding from millions of predicted structuresChloe Hsu, Robert Verkuil, Jason Liu, Zeming Lin et al.ICML 2022 · 560 citations
- Riemannian Score-Based Generative ModellingValentin De Bortoli, Emile Mathieu, Michael J. Hutchinson, James Thornton et al.NeurIPS 2022 · 306 citations
- Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-DesignAndrew Campbell, Jason Yim, Regina Barzilay, Tom Rainforth et al.ICML 2024 · 283 citations
- Analog Bits: Generating Discrete Data using Diffusion Models with Self-ConditioningTing Chen, Ruixiang Zhang, Geoffrey E. HintonICLR 2023 · 80 citations
Related papers
- Pallatom-Ligand: an All-Atom Diffusion Model for Designing Ligand-Binding ProteinsHaochen Wang, Qianyi Wang, Rui Ma, Jiawei Guan et al.ICLR 2026
- Sequence-Augmented SE(3)-Flow Matching For Conditional Protein GenerationGuillaume Huguet, James Vuckovic, Kilian Fatras, Eric Thibodeau-Laufer et al.NeurIPS 2024 · 32 citations
- SLAE: Strictly Local All-atom Environment for Protein RepresentationYilin Chen, Tianyu Lu, Cizhang Zhao, Hannah Wayment-Steele et al.ICML 2026
- ProAR: Probabilistic Autoregressive Modeling for Molecular DynamicsKaiwen Cheng, Yutian Liu, Zhiwei Nie, Mujie Lin et al.AAAI 2026
- AbDiffuser: full-atom generation of in-vitro functioning antibodiesKarolis Martinkus, Jan Ludwiczak, Wei-Ching Liang, Julien Lafrance-Vanasse et al.NeurIPS 2023 · 79 citations
