Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jan 15, 2026

Constructing Cyclic Peptides Using an On-Tether Sulfonium Center
07:11

Constructing Cyclic Peptides Using an On-Tether Sulfonium Center

Published on: September 28, 2022

3.1K

HighMPNN: A Graph Neural Network Approach for Structure-Constrained Cyclic Peptide Sequence Design.

Wen Xu, Chengyun Zhang, Tianfeng Shang

    IEEE Journal of Biomedical and Health Informatics
    |October 13, 2025
    PubMed
    Summary

    HighMPNN is a new graph neural network model designed for cyclic peptide sequence design. It improves structural accuracy and sequence recovery, accelerating the discovery of novel peptide therapeutics.

    Related Concept Videos

    You might also read

    Related Articles

    Articles linked to this work by shared authors, journal, and citation graph.

    Sort by
    Same author

    HighDB: a structure-annotated cyclic peptide database for comparative analysis, template retrieval, and design-oriented applications.

    Journal of computer-aided molecular design·2026
    Same author

    HFGuidedDesign: <i>de novo</i> design of cyclic peptide binders <i>via</i> structure-guided discrete diffusion.

    Chemical science·2026
    Same author

    MIFNDRA: an innovative knowledge-enhanced multimodal fusion and graph learning framework for predicting drug resistance-related ncRNAs.

    Briefings in bioinformatics·2026
    Same author

    HighRes_Builder: improved access and modeling of noncanonical residues for protein structure prediction.

    Briefings in bioinformatics·2026
    Same author

    AI-Designed Cyclic Peptides Enable Controllable Modulation of the CD28 Immune Checkpoint.

    Advanced science (Weinheim, Baden-Wurttemberg, Germany)·2026
    Same author

    EZPro-Multi: Contrastive Learning-Enhanced Multi-property Prediction for Enzyme Engineering.

    Journal of chemical theory and computation·2026

    Area of Science:

    • Computational biology
    • Peptide chemistry
    • Drug discovery

    Background:

    • Cyclic peptides show therapeutic potential but designing their sequences is challenging.
    • Existing deep learning models are not optimized for the topological constraints of cyclic peptides.

    Purpose of the Study:

    • To develop a novel graph neural network (GNN) model, HighMPNN, for de novo cyclic peptide sequence design.
    • To address the limitations of current models in capturing cyclic peptide structural constraints.

    Main Methods:

    • Developed HighMPNN, a GNN model integrating explicit structural constraints for cyclic backbones.
    • Employed a combined loss function including cross-entropy and Frame Aligned Point Error (FAPE) for sequence generation and structural accuracy.
    • Evaluated model performance using sequence recovery rate and Cα root-mean-square deviation (RMSD_Cα).

    More Related Videos

    Constructing Thioether/Vinyl Sulfide-tethered Helical Peptides Via Photo-induced Thiol-ene/yne Hydrothiolation
    11:09

    Constructing Thioether/Vinyl Sulfide-tethered Helical Peptides Via Photo-induced Thiol-ene/yne Hydrothiolation

    Published on: August 1, 2018

    11.2K
    Synthesis of Information-bearing Peptoids and their Sequence-directed Dynamic Covalent Self-assembly
    09:34

    Synthesis of Information-bearing Peptoids and their Sequence-directed Dynamic Covalent Self-assembly

    Published on: February 6, 2020

    7.9K

    Related Experiment Videos

    Last Updated: Jan 15, 2026

    Constructing Cyclic Peptides Using an On-Tether Sulfonium Center
    07:11

    Constructing Cyclic Peptides Using an On-Tether Sulfonium Center

    Published on: September 28, 2022

    3.1K
    Constructing Thioether/Vinyl Sulfide-tethered Helical Peptides Via Photo-induced Thiol-ene/yne Hydrothiolation
    11:09

    Constructing Thioether/Vinyl Sulfide-tethered Helical Peptides Via Photo-induced Thiol-ene/yne Hydrothiolation

    Published on: August 1, 2018

    11.2K
    Synthesis of Information-bearing Peptoids and their Sequence-directed Dynamic Covalent Self-assembly
    09:34

    Synthesis of Information-bearing Peptoids and their Sequence-directed Dynamic Covalent Self-assembly

    Published on: February 6, 2020

    7.9K

    Main Results:

    • HighMPNN achieved a 63.95% average sequence recovery rate.
    • The model demonstrated an average Cα root-mean-square deviation (RMSD_Cα) of 1.413 Å, indicating high structural consistency.
    • HighMPNN outperformed baseline models in both sequence recovery and structural accuracy.

    Conclusions:

    • HighMPNN effectively designs cyclic peptide sequences with high structural fidelity.
    • The model's ability to learn sequence patterns while respecting geometric constraints accelerates cyclic peptide discovery.
    • Future work will extend HighMPNN to non-canonical residues and diverse scaffolds for broader therapeutic applications.