Sheng Xu

Sheng Xu

PhD Student in Computational Biology

AI for Molecular Biology • Computational Virology • Language Models

Get in Touch Publications Google Scholar

Recent News

May 2025 — FoldBench preprint released on bioRxiv, introducing a comprehensive all-atom benchmark for structure prediction methods.
May 2025 — CRAFTS published in Science. 🎉
Jan 2025 — π‑PrimeNovo published in Nature Communications. 🎉
Apr 2024 — SSBlazer published in Genome Biology. Many thanks to all collaborators! 🎉

Research Interests

Virus Discovery & Host Prediction

Developing foundation models for viral genome analysis, integrating structure, ecology, and host factors. Building scalable pipelines for metagenomic mining and zoonotic risk assessment.

Viral Language Models Zoonosis Prediction Metagenomics Risk Assessment

Structure Prediction & Design

Creating comprehensive benchmarks for all-atom structure prediction models. Developing generative approaches for protein design, antibody engineering, and molecular therapeutics with robust uncertainty quantification.

FoldBench Antibody Design Diffusion Models Uncertainty Quantification

Experience

Education

Fudan University
2023 – present
PhD in Biomedical Engineering
The Chinese University of Hong Kong
2021 – 2022
MSc in Computer Science
East China Normal University
2017 – 2021
BEng in Software Engineering

Employment

Shanghai Artificial Intelligence Laboratory
2022 – present
Research Intern

Publications

  1. Curriculum Learning for Biological Sequence Prediction: The Case of De Novo Peptide Sequencing
    X. Zhang, J. Wei, Z. Qiu, S. Xu, et al.
    ICML 2025
    Paper
  2. Universal Biological Sequence Reranking for Improved De Novo Peptide Sequencing
    Z. Qiu, J. Wei, X. Zhang, S. Xu, et al.
    ICML 2025
    Paper Code
  3. Cryo‑EM reveals mechanisms of natural RNA multivalency
    L. Wang*, J. Xie*, T. Gong*, H. Wu*, Y. Tu*, X. Peng*, S. Shang*, X. Jia*, H. Ma*, J. Zou*, S. Xu*, et al.
    Science, 2025
    Paper
  4. Fast, sensitive detection of protein homologs using deep dense retrieval
    L. Hong, Z. Hu, S. Sun, X. Tang, J. Wang, Q. Tan, L. Zheng, S. Wang, S. Xu, et al.
    Nature Biotechnology, 2025
    Paper Code
  5. π‑PrimeNovo: an accurate and efficient non‑autoregressive deep learning model for de novo peptide sequencing
    X. Zhang*, T. Ling*, Z. Jin*, S. Xu*, et al.
    Nature Communications, 2025
    Paper Code
  6. SSBlazer: a genome-wide nucleotide-resolution model for predicting single-strand DNA breaks
    S. Xu*, J. Wei*, et al.
    Genome Biology, 2024
    Paper
  7. Beacon: Benchmark for comprehensive RNA tasks and language models
    Y. Ren, Z. Chen, L. Qiao, H. Jing, Y. Cai, S. Xu, et al.
    NeurIPS 2024
    Paper
  8. Accurate prediction of antibody function and structure using a bio‑inspired antibody language model
    H. Jing, Z. Gao, S. Xu, et al.
    Briefings in Bioinformatics, 2024
    Paper Code
  9. CrossBind: Collaborative cross‑modal identification of protein nucleic‑acid‑binding residues
    L. Jing*, S. Xu*, et al.
    AAAI 2024
    Paper Code
  10. ContraNovo: a contrastive learning approach to enhance de novo peptide sequencing
    Z. Jin*, S. Xu*, et al.
    AAAI 2024
    Paper Code
  11. AF2‑Mutation: adversarial sequence mutations against AlphaFold2 on protein tertiary structure prediction
    Z. Yuan, T. Shen, S. Xu, et al.
    Acta Materia Medica, 2024
    Paper
  12. AcrNET: predicting anti‑CRISPR with deep learning
    Y. Li, Y. Wei, S. Xu, et al.
    Bioinformatics, 2023
    Paper

Preprints & Working Papers

  1. A genome-scale language model uncovers animal viruses with zoonotic potential
    S. Xu, et al.
    Submitted, 2025
  2. Fitness aligned structural modeling enables scalable virtual screening with AuroBind
    Z. Zhang, J. Rao, J. Zhong, W. Bai, D. Wang, S. Ning, L. Qiao, S. Xu, et al.
    arXiv, 2025
    Paper
  3. FoldBench: An all‑atom benchmark for biomolecular structure prediction
    S. Xu*, Q. Feng*, L. Qiao, H. Wu, T. Shen, Y. Cheng, S. Zheng, S. Sun
    bioRxiv, 2025
    Paper
  4. RTM‑align: an improved RNA alignment tool with enhanced short sequence performance via post‑standardization and fragment alignment
    Z. Qiu*, S. Xu*, J. Wei, T. Shen, S. Sun
    bioRxiv, 2024
    Paper

Contact & Collaboration

I am always open to discussions about research collaborations, particularly in the areas of virus discovery, protein design, and protein/genome language models. Feel free to reach out if you're interested in working together or have questions about my research.

Email: shengxu@link.cuhk.edu.hk • ORCID: 0000-0002-6507-9122

GitHub Google Scholar LinkedIn