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

Is cross-validation valid for small-sample microarray classification?

Ulisses M Braga-Neto1, Edward R Dougherty

  • 1Section of Clinical Cancer Genetics, University of Texas MD Anderson Cancer Center, Houston, TX, USA.

Bioinformatics (Oxford, England)
|February 13, 2004
PubMed
Summary
This summary is machine-generated.

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

Pathway-based analyses of gene expression profiles at low doses of ionizing radiation.

Frontiers in bioinformatics·2024
Same author

Optimal decision-making in high-throughput virtual screening pipelines.

Patterns (New York, N.Y.)·2023
Same author

Knowledge-driven learning, optimization, and experimental design under uncertainty for materials discovery.

Patterns (New York, N.Y.)·2023
Same author

Sensitivity Analysis of Genome-Scale Metabolic Flux Prediction.

Journal of computational biology : a journal of computational molecular cell biology·2023
Same author

Short-chain fatty acid production in accessible and inaccessible body pools as assessed by novel stable tracer pulse approach is reduced by aging independent of presence of COPD.

Metabolism: clinical and experimental·2023
Same author

Probabilistic boolean networks predict transcription factor targets to induce transdifferentiation.

iScience·2022
Same journal

3DICE: Interpretable 3D Cross-Modal Learning for Drug-Target Interaction Prediction and Large-Scale Drug Discovery.

Bioinformatics (Oxford, England)·2026
Same journal

KASSPer: Kinase Active Site Structure Prediction using Protein and Ligand Language Models and Its Application to Virtual Screening.

Bioinformatics (Oxford, England)·2026
Same journal

IDR searcher: a search engine solution for public image resources.

Bioinformatics (Oxford, England)·2026
Same journal

KCFtools: Rapid alignment-free method for introgression screening and GWAS using k-mer profiles.

Bioinformatics (Oxford, England)·2026
Same journal

Meta2DB: Curated shotgun metagenomic feature sets and metadata for health state prediction.

Bioinformatics (Oxford, England)·2026
Same journal

conMItion: an R package adjusting confounding factors for associations in multi-omics.

Bioinformatics (Oxford, England)·2026
See all related articles

For small sample sizes in microarray classification, cross-validation error estimation is less biased but highly variable. Bootstrap methods offer better precision but can increase bias and computational cost.

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Statistical Learning

Background:

  • Microarray classification often uses small sample sizes for classifier design and error estimation.
  • Cross-validation is the predominant method for error estimation in these studies.
  • Understanding cross-validation's behavior with limited data is crucial.

Purpose of the Study:

  • To compare the performance of cross-validation, resubstitution, and bootstrap error estimation methods.
  • To evaluate these methods under conditions of very small sample sizes.
  • To assess bias and precision of different error estimation techniques in microarray classification.

Main Methods:

  • An extensive simulation study was conducted.
  • Compared linear discriminant analysis, 3-nearest-neighbor, and decision trees (CART).

Related Experiment Videos

  • Utilized synthetic and real breast-cancer patient data.

    • Main Results:

    • Cross-validation shows lower bias than resubstitution but exhibits high variance, leading to unreliable estimates for small samples.
    • Bootstrap methods improve variance but can increase bias and computational expense.
    • Resubstitution demonstrates the highest bias among the methods.

    Conclusions:

    • Cross-validation's high variance limits its reliability for small sample sizes in microarray classification.
    • Bootstrap methods present a trade-off between variance reduction and potential bias increase.
    • Careful consideration of error estimation methods is necessary when dealing with limited data in bioinformatics.