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: Jun 19, 2026

DeepOmicsAE: Representing Signaling Modules in Alzheimer's Disease with Deep Learning Analysis of Proteomics, Metabolomics, and Clinical Data
09:47

DeepOmicsAE: Representing Signaling Modules in Alzheimer's Disease with Deep Learning Analysis of Proteomics, Metabolomics, and Clinical Data

Published on: December 15, 2023

Learning High-Resolution Protein Embeddings from Multimodal Data via Self-Supervised Integration.

Yong-Jia Liang1,2,3, Qian-Yi Wang1,2,3, Qian Zhou1,2,3

  • 1School of Biomedical Engineering, Southern Medical University, Guangzhou 510515, China.

Journal of Chemical Information and Modeling
|June 18, 2026
PubMed
Summary

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

Three-Dimensional Visualisation of Blood Vessels in Human Gliomas Using Tissue Clearing and Deep Learning.

Neuropathology and applied neurobiology·2025
Same author

Knowledge-enhanced protein subcellular localization prediction from 3D fluorescence microscope images.

Bioinformatics (Oxford, England)·2025
Same author

Deep learning-based vessel extraction in 3D confocal microscope images of cleared human glioma tissues.

Biomedical optics express·2024
Same author

Automated identification of protein expression intensity and classification of protein cellular locations in mouse brain regions from immunofluorescence images.

Medical & biological engineering & computing·2023
Same author

Automatic recognition of protein subcellular location patterns in single cells from immunofluorescence images based on deep learning.

Briefings in bioinformatics·2022
Same author

GraphLoc: a graph neural network model for predicting protein subcellular localization from immunohistochemistry images.

Bioinformatics (Oxford, England)·2022

This study introduces self-SSGI, a novel multimodal self-supervised method for learning protein embeddings. It effectively integrates diverse protein data, significantly improving performance on various protein annotation tasks.

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Structural Biology

Background:

  • Vast amounts of multimodal protein data (sequences, structures, GO annotations, images) are available.
  • Experimentally validated protein function annotations are limited.
  • Accurate, low-dimensional protein representations are crucial for machine learning-based annotation.

Purpose of the Study:

  • To develop a multimodal self-supervised learning method for high-resolution protein embeddings.
  • To address the limitations of unimodal deep learning approaches in protein representation.
  • To enhance protein annotation by integrating diverse data modalities.

Main Methods:

  • Developed self-SSGI, a multimodal self-supervised method integrating sequence, structure, GO annotations, and images.

Related Experiment Videos

Last Updated: Jun 19, 2026

DeepOmicsAE: Representing Signaling Modules in Alzheimer's Disease with Deep Learning Analysis of Proteomics, Metabolomics, and Clinical Data
09:47

DeepOmicsAE: Representing Signaling Modules in Alzheimer's Disease with Deep Learning Analysis of Proteomics, Metabolomics, and Clinical Data

Published on: December 15, 2023

  • Employed a joint masked reconstruction strategy for amino acid-level features.
  • Utilized contrastive learning for protein-level features and a cross-attention module for multimodal fusion.

    • Main Results:

    • self-SSGI effectively integrates multimodal protein data, learning enhanced protein embeddings.
    • The method achieved superior performance on protein subcellular localization, molecular function prediction, and protein-protein interaction inference.
    • Results demonstrated strong performance on both training and external datasets, surpassing state-of-the-art methods.

    Conclusions:

    • self-SSGI provides a powerful tool for generating high-quality protein representations.
    • The method's ability to integrate multimodal data enhances downstream bioinformatics tasks.
    • This work facilitates further computational research and protein annotation in bioinformatics.