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

Comparison and evaluation of methods for generating differentially expressed gene lists from microarray data.

Ian B Jeffery1, Desmond G Higgins, Aedín C Culhane

  • 1Bioinformatics, Conway Institute, University College Dublin, Belfield, Dublin 4, Ireland. Ian.Jeffery@ucd.ie

BMC Bioinformatics
|July 29, 2006
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

Convergence of machine learning and genomics for precision oncology.

Nature reviews. Cancer·2026
Same author

Learning and teaching biological data science in the Bioconductor community.

PLoS computational biology·2025
Same author

Learning and teaching biological data science in the Bioconductor community.

ArXiv·2025
Same author

demuxSNP: supervised demultiplexing single-cell RNA sequencing using cell hashing and SNPs.

GigaScience·2024
Same author

Dynamics of the gut microbiome in individuals at risk of rheumatoid arthritis: a cross-sectional and longitudinal observational study.

Annals of the rheumatic diseases·2024
Same author

The 1+Million Genomes Minimal Dataset for Cancer.

Nature genetics·2024
Same journal

Covariance decomposition for distance based species tree estimation.

BMC bioinformatics·2026
Same journal

SNPio: a Python interface for population genomic data processing.

BMC bioinformatics·2026
Same journal

SpaHNR: a spatial domain identification method via sparse attention-based hierarchical node representation and multi-view contrastive learning.

BMC bioinformatics·2026
Same journal

OpenIMC: an open-source platform for analyzing single-cell and spatial proteomics by imaging mass cytometry.

BMC bioinformatics·2026
Same journal

NAP: an open source pipeline for cross-domain microbiome profiling using Nanopore sequencing-derived amplicon data.

BMC bioinformatics·2026
Same journal

SurvGME: an R package for survival analysis with graphical and measurement error models.

BMC bioinformatics·2026
See all related articles

Comparing feature selection methods for microarray data reveals significant differences in gene lists and classification performance. The best method depends on dataset characteristics like noise and sample size, with Area Under the ROC Curve, Rank Products, and Empirical Bayes t-statistic showing specific strengths.

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Gene Expression Analysis

Background:

  • Numerous feature selection methods exist for identifying differentially expressed genes in microarray data.
  • Common methods like fold change, t-statistic, and moderated t-statistics often yield dissimilar gene lists.
  • Few direct comparisons of these feature selection methods have been conducted.

Purpose of the Study:

  • To empirically compare commonly used feature selection methods for microarray data.
  • To evaluate the gene lists produced by these methods and their performance in class prediction.
  • To provide recommendations for selecting appropriate feature selection methods based on dataset characteristics.

Main Methods:

  • Applied ten feature selection methods (SAM, ANOVA, Empirical Bayes t-statistic, template matching, maxT, BGA, ROC AUC, Welch t-statistic, fold change, rank products) to nine binary microarray datasets.

Related Experiment Videos

  • Compared the gene lists generated by each method for agreement.
  • Evaluated the class prediction efficiency of each gene list using cross-validation and four supervised classifiers.

    • Main Results:

    • Little agreement was observed in gene lists produced by different methods, with only 8-21% of genes in common across all ten methods.
    • Classification performance varied significantly based on the feature selection method and dataset characteristics.
    • Area Under the ROC Curve (ROC AUC) performed well on datasets with low noise and large sample sizes.

    Conclusions:

    • The choice of feature selection method, number of genes, sample size, and dataset noise substantially influence classification success.
    • Rank products are effective for datasets with few samples or high noise.
    • Empirical Bayes t-statistic demonstrated robust performance across various sample sizes.