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

Penalized binary regression for gene expression profiling.

Michael G Schimek1

  • 1Medical University of Graz, Institute for Medical Informatics, Statistics and Documentation, Auenbruggerplatz 2, 8036 Graz, Austria. michael.schimek@meduni-graz.at

Methods of Information in Medicine
|February 11, 2005
PubMed
Summary

Penalization enhances binary regression for microarray analysis, enabling accurate classification of biological samples. Both frequentist and Bayesian methods offer robust solutions for this bioinformatics challenge.

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

Parea: Multi-view ensemble clustering for cancer subtype discovery.

Journal of biomedical informatics·2023
Same author

Robust Random Forest-Based All-Relevant Feature Ranks for Trustworthy AI.

Studies in health technology and informatics·2022
Same author

A hierarchical clustering and data fusion approach for disease subtype discovery.

Journal of biomedical informatics·2020
Same author

Genome scans for selection and introgression based on k-nearest neighbour techniques.

Molecular ecology resources·2020
Same author

Publisher Correction: Inference of transcription factor binding from cell-free DNA enables tumor subtype prediction and early detection.

Nature communications·2020
Same author

Inference of transcription factor binding from cell-free DNA enables tumor subtype prediction and early detection.

Nature communications·2019

Area of Science:

  • Bioinformatics
  • Statistical Genetics
  • Computational Biology

Background:

  • Microarray analysis frequently involves classifying biological samples into distinct categories.
  • Accurate classification requires robust statistical methods that can handle high-dimensional data and small sample sizes, common in gene expression studies.

Purpose of the Study:

  • To develop and evaluate a statistical procedure for classifying biological samples based on gene expression levels.
  • To adapt binary regression techniques for effective microarray data analysis, addressing limitations of standard methods.

Main Methods:

  • Utilized binary regression, specifically penalized logit and Bayesian probit models, to address high-dimensionality and small sample size challenges.
  • Introduced frequentist and Bayesian penalization concepts for improved model stability and predictive performance.

Related Experiment Videos

  • Explored the role of cross-validation in regularization and feature selection for robust classification.

    • Main Results:

    • Penalization significantly improves the suitability of classical binary regression for microarray data analysis.
    • Demonstrated the application of penalized logit and Bayesian probit regression on a benchmark dataset.
    • Compared the performance of penalized methods against decision tree algorithms, showing competitive results.

    Conclusions:

    • Both frequentist and Bayesian penalization approaches are effective for microarray classification.
    • Identified method-specific differences, with the Bayesian approach providing posterior probabilities to quantify assumption-induced bias.