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 Videos

Hierarchical multi-label prediction of gene function.

Zafer Barutcuoglu1, Robert E Schapire, Olga G Troyanskaya

  • 1Department of Computer Science, Princeton University, 35 Olden Street, Princeton, NJ 08544, USA.

Bioinformatics (Oxford, England)
|January 18, 2006
PubMed
Summary

This study introduces a Bayesian framework to improve gene function prediction by leveraging the Gene Ontology (GO) hierarchy. The new method ensures consistent predictions and enhances accuracy for unknown genes.

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

Single-cell epigenetic landscape, microenvironment interactions, and gene regulatory modules of non-functioning pituitary adenomas.

Cell systems·2026
Same author

Shared multicellular injury programs of acute and chronic kidney disease enable mechanistic patient stratification.

medRxiv : the preprint server for health sciences·2026
Same author

Single-cell profiling reveals epithelial and immune responses in BK polyomavirus-infected human kidney biopsies.

JCI insight·2026
Same author

Multi-modal tissue-aware graph neural network for <i>in silico</i> genetic discovery.

bioRxiv : the preprint server for biology·2026
Same author

Gene-centered representation of coding and regulatory variation enables outcome prediction.

bioRxiv : the preprint server for biology·2026
Same author

Variant-resolved prediction of context-specific isoform variation with a graph-based attention model.

Cell genomics·2026

Area of Science:

  • Genomics
  • Bioinformatics
  • Computational Biology

Background:

  • Predicting gene function is crucial in functional genomics.
  • Existing methods often use independent classifiers, ignoring functional class taxonomy structures like the Gene Ontology (GO).
  • This can lead to inconsistent predictions within the GO hierarchy.

Purpose of the Study:

  • To develop a Bayesian framework that incorporates functional taxonomy constraints for improved gene function prediction.
  • To resolve hierarchical inconsistencies in predictions generated by independent classifiers.
  • To enhance the specificity of gene function predictions by utilizing the entire classifier hierarchy.

Main Methods:

  • Developed a Bayesian framework to combine multiple classifiers.

Related Experiment Videos

  • Utilized a hierarchy of support vector machine (SVM) classifiers trained on diverse data types.
  • Integrated predictions within the framework to achieve the most probable consistent set.

    • Main Results:

    • The Bayesian framework improved predictions for 93 out of 105 GO nodes in experiments.
    • The method provides implicit calibration of SVM margin outputs to probabilities.
    • Function predictions were made for multiple proteins, with experimental confirmation for mitosis-related proteins.

    Conclusions:

    • The Bayesian framework effectively leverages hierarchical structures for more accurate gene function prediction.
    • This approach resolves inconsistencies and enhances prediction specificity.
    • The method offers a robust way to predict and validate gene functions.