Lune

ICML2025Top-tier venue

CFP-Gen: Combinatorial Functional Protein Generation via Diffusion Language Models

Junbo Yin, Chao Zha, Wenjia He, Chencheng Xu, Xin Gao

Abstract

Existing PLMs generate protein sequences based on a single-condition constraint from a specific modality, struggling to simultaneously satisfy multiple constraints across different modalities. In this work, we introduce CFP-GEN, a novel diffusion language model for Combinatorial Functional Protein GENeration. CFP-GEN facilitates the de novo protein design by integrating multimodal conditions with functional, sequence, and structural constraints. Specifically, an Annotation-Guided Feature Modulation (AGFM) module is introduced to dynamically adjust the protein feature distribution based on composable functional annotations, e.g., GO terms, IPR domains and EC numbers. Meanwhile, the Residue-Controlled Functional Encoding (RCFE) module captures residue-wise interaction to ensure more precise control. Additionally, off-the-shelf 3D structure encoders can be seamlessly integrated to impose geometric constraints. We demonstrate that CFP-GEN enables high-throughput generation of novel proteins with functionality comparable to natural proteins, while achieving a high success rate in designing multifunctional proteins.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

3 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Diffusion Language Models Are Versatile Protein Learners cites Structured Denoising Diffusion Models in Discrete State-Spaces
  • Diffusion Language Models Are Versatile Protein Learners cites Learning inverse folding from millions of predicted structures
  • Diffusion Language Models Are Versatile Protein Learners cites Structure-informed Language Models Are Protein Designers
  • MMCP-GEN: A Modality-Extensible Diffusion Language Model for Conditional Protein Sequence Generation cites Diffusion Language Models Are Versatile Protein Learners
  • MMCP-GEN: A Modality-Extensible Diffusion Language Model for Conditional Protein Sequence Generation cites Learning inverse folding from millions of predicted structures
  • MMCP-GEN: A Modality-Extensible Diffusion Language Model for Conditional Protein Sequence Generation cites Structure-informed Language Models Are Protein Designers
  • PDFBench: A Benchmark for De Novo Protein Design from Function cites Diffusion Language Models Are Versatile Protein Learners
  • PDFBench: A Benchmark for De Novo Protein Design from Function cites Learning inverse folding from millions of predicted structures
  • Co-Generative De Novo Functional Protein Design cites Diffusion Language Models Are Versatile Protein Learners
  • Co-Generative De Novo Functional Protein Design cites Structured Denoising Diffusion Models in Discrete State-Spaces
  • Structure-informed Language Models Are Protein Designers cites Learning inverse folding from millions of predicted structures
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers3

  1. PDFBench: A Benchmark for De Novo Protein Design from FunctionJiahao Kuang, Nuowei Liu, Changzhi Sun, Jie Wang et al.ICML 2026 · 10 citations
  2. MMCP-GEN: A Modality-Extensible Diffusion Language Model for Conditional Protein Sequence GenerationZeyu An, Wanyu Lin, Feng Tan, Shujun WangCVPR 2026
  3. Co-Generative De Novo Functional Protein DesignXinRui Chen, YIZHEN LUO, Siqi Fan, Zaiqing NieICML 2026

Builds on6

  1. Adding Conditional Control to Text-to-Image Diffusion ModelsLvmin Zhang, Anyi Rao, Maneesh AgrawalaICCV 2023 · 6,759 citations
  2. Structured Denoising Diffusion Models in Discrete State-SpacesJacob Austin, Daniel D. Johnson, Jonathan Ho, Daniel Tarlow et al.NeurIPS 2021 · 2,256 citations
  3. Learning inverse folding from millions of predicted structuresChloe Hsu, Robert Verkuil, Jason Liu, Zeming Lin et al.ICML 2022 · 560 citations
  4. Structure-informed Language Models Are Protein DesignersZaixiang Zheng, Yifan Deng, Dongyu Xue, Yi Zhou et al.ICML 2023 · 130 citations
  5. Diffusion Language Models Are Versatile Protein LearnersXinyou Wang, Zaixiang Zheng, Fei Ye, Dongyu Xue et al.ICML 2024 · 113 citations

Related papers

  1. Proteo-R1: Reasoning Foundation Models for De Novo Protein DesignFang Wu, Weihao Xuan, Heli Qi, Hanqun CAO et al.ICML 2026 · 5 citations
  2. DPLM-2: A Multimodal Diffusion Protein Language ModelXinyou Wang, Zaixiang Zheng, Fei Ye, Dongyu Xue et al.ICLR 2025
  3. Annotation-guided Protein Design with Multi-Level Domain AlignmentChaohao Yuan, Songyou Li, Geyan Ye, Yikun Zhang et al.KDD 2025 · 2 citations
  4. CAGenMol: Condition-Aware Diffusion Language Model for Goal-Directed Molecular GenerationYanting Li, Zhuoyang Jiang, Enyan Dai, Lei Wang et al.ACL Findings 2026
  5. DS-ProGen: A Dual-Structure Deep Language Model for Functional Protein DesignYanting Li, Zikang Wang, Jiyue Jiang, Ziqian Lin et al.AAAI 2026
Dusk over the sea between two cliffs drawn in fine vertical lines