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 Concept Videos

Multi-pass Transmembrane Proteins and β-barrels01:09

Multi-pass Transmembrane Proteins and β-barrels

In multi-pass transmembrane proteins, the polypeptide chain crosses the membrane more than once. The transmembrane polypeptide chain either forms an α-helix or β-strand structure. α-Helix containing multi-pass transmembrane proteins are ubiquitous, whereas β-strand containing ones are mainly found in gram-negative bacteria, mitochondria, and chloroplasts.
α-Helix containing multi-pass transmembrane proteins
Multi-pass transmembrane proteins such as G-protein-linked receptors (GPCRs) and...
Two-Dimensional Microscopy in Microbiology01:29

Two-Dimensional Microscopy in Microbiology

Two-dimensional (2D) microscopy encompasses a range of optical techniques that capture images within a single focal plane, offering detailed representations of microscopic structures. These techniques are essential in biological and medical research, enabling the visualization of cellular and subcellular structures with different levels of contrast and specificity.There are several major types of 2D microscopy, each with strengths and applications.Bright-Field MicroscopyBright-field microscopy...
Multiple Bar Graph01:07

Multiple Bar Graph

As the name suggests, a multiple bar graph is the same as a bar graph but has multiple bars to depict relationships between different data values. One can include as many parameters as possible. However, each parameter must have the same unit of measurement.
Each bar or column in the multiple bar graph represents a data value. These graphs are used primarily in interrelating two or more sets of data. The categories of different kinds of data are listed along the horizontal or x-axis, whereas...

You might also read

Related Articles

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

Sort by
Same author

scCCVGBen for benchmarking of single-cell representation learning anchored on a centroid-coupled variational graph attention autoencoder across scRNA-seq and scATAC-seq.

Frontiers in genetics·2026
Same author

LAIOR: a hyperbolic neural ODE variational framework for interpretable single-cell manifold learning and trajectory inference.

Frontiers in genetics·2026
Same author

Obesity-induced oleic acid metabolic dysregulation may exacerbate osteoarthritis through the degradation of SOX9.

Cellular and molecular life sciences : CMLS·2026
Same author

Bioinspired micro/nanofibers interlocking for high-strength self-bonded bamboo material.

International journal of biological macromolecules·2026
Same author

Institutional incentives and faculty cooperation in scholarship of teaching and learning: a scenario-based study in medical education.

BMC medical education·2026
Same author

Direct Photolithography of Fluorescent Copper (I) Iodide Cluster Microarrays.

Small methods·2026

Related Experiment Video

Updated: Jul 12, 2026

A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells
12:49

A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells

Published on: September 28, 2019

GAHIB: graph attention VAE with a hyperbolic information bottleneck for biologically structured single-cell

Zeyu Fu1, Jiawei Fu2, Xiaoxia Wang3

  • 1State Key Laboratory of Trauma and Chemical Poisoning, Institute of Combined Injury, Chongqing Engineering Research Center for Nanomedicine, College of Preventive Medicine, Army Medical University, Chongqing, China.

Frontiers in Genetics
|July 11, 2026
PubMed
Summary

GAHIB, a novel graph attention VAE with hyperbolic information bottleneck, integrates multiple biases in single-cell RNA sequencing. It offers a balanced performance across benchmarks and provides a geometrically aware representation for biological interpretation.

Keywords:
graph attention networkhyperbolic geometryinformation bottleneckrepresentation learningsingle-cell RNA-seqvariational autoencoder

More Related Videos

Three-dimensional Imaging of Bacterial Cells for Accurate Cellular Representations and Precise Protein Localization
06:33

Three-dimensional Imaging of Bacterial Cells for Accurate Cellular Representations and Precise Protein Localization

Published on: October 29, 2019

Related Experiment Videos

Last Updated: Jul 12, 2026

A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells
12:49

A Method for 3D Reconstruction and Virtual Reality Analysis of Glial and Neuronal Cells

Published on: September 28, 2019

Three-dimensional Imaging of Bacterial Cells for Accurate Cellular Representations and Precise Protein Localization
06:33

Three-dimensional Imaging of Bacterial Cells for Accurate Cellular Representations and Precise Protein Localization

Published on: October 29, 2019

Area of Science:

  • Computational Biology
  • Genomics
  • Machine Learning

Background:

  • Current single-cell RNA sequencing (scRNA-seq) variational autoencoders (VAEs) often focus on individual aspects like graph structure, hyperbolic geometry, or compression.
  • These biases are rarely evaluated together in a unified model, limiting comprehensive understanding of their impact.

Purpose of the Study:

  • To introduce GAHIB (graph attention VAE with a hyperbolic information bottleneck), a novel model that integrates graph attention, a 2D information bottleneck, and hyperbolic geometry.
  • To evaluate GAHIB's performance across a broad benchmark and in specific biological contexts, comparing it with existing methods.

Main Methods:

  • GAHIB combines a graph-attention encoder, a 2D information bottleneck, and a Lorentz-hyperbolic geometry loss.
  • Model evaluation was conducted in two tiers: a 53-dataset proxy-label benchmark and curated-label analyses on annotated systems.
  • Performance was assessed using clustering, projection quality, latent structure metrics, and biological interpretation via marker enrichment.

Main Results:

  • GAHIB demonstrated a balanced performance across clustering, projection quality, and latent structure metrics on the proxy benchmark.
  • Comparisons showed scVI and scDHMap remain competitive in specific metrics (NMI/ARI, DRE-UMAP).
  • Biological interpretation on curated systems showed context-dependent signals, with GAHIB supporting lineage structure in muscle atlases and scVI excelling in T-cell subtype classification.

Conclusions:

  • GAHIB offers a complementary, geometrically aware representation for single-cell data, rather than a direct replacement for clustering methods.
  • The model shows practical runtime and stability, but limitations include context-specific dropout evidence and lack of complete manual curation.
  • Further research is needed to fully explore its potential in diverse biological applications.