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

Importance of data structure in comparing two dimension reduction methods for classification of microarray gene

Caroline Truntzer1, Catherine Mercier, Jacques Estève

  • 1CNRS, UMR 5558--Equipe Biostatistique Santé, Villeurbanne, France. caroline.truntzer@chu-lyon.fr

BMC Bioinformatics
|March 16, 2007
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

An expression signature of 19 human endogenous retroviruses identifies immunogenic luminal breast cancers likely to respond to immunotherapy.

Frontiers in oncology·2026
Same author

Deficits in Cognitive-Motor Control of the Ipsilesional Upper Limb in Subacute Stroke Assessed Using a Robotic Exoskeleton: A Longitudinal Study.

Brain sciences·2026
Same author

IL-10 secretion by CD8 T cells orchestrates early NK cell recruitment and antitumor immunity after anti-VEGFR-2/PD-1 therapy.

Journal for immunotherapy of cancer·2026
Same author

Pooled analysis of 2 clinical trials of first-line chemoimmunotherapy for metastatic microsatellite stable colorectal cancer MEDITREME and METIMMOX studies.

Oncoimmunology·2026
Same author

Changes in Unimanual and Bimanual Upper Extremity Use During the Subacute Phase Post-Stroke Assessed in Supervised and Unsupervised Contexts.

Neurorehabilitation and neural repair·2026
Same author

"I See What Others Don't See When They Look at Me": A Qualitative Study of Body Perception Disturbances in Chronic Non-Cancer Pain.

Journal of pain research·2026

This study compares gene expression analysis methods, finding Discriminant Analysis superior due to its ability to handle data structure. Understanding data structure is key to selecting the best method for accurate gene classification and prognosis.

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Statistical Genetics

Background:

  • Microarray technology has enabled numerous gene classification and prognosis methods.
  • Some methods, despite different names, share similar underlying approaches.
  • This study investigates how gene expression variance structure impacts method performance.

Purpose of the Study:

  • To evaluate the influence of gene expression variance structure on classification and prognosis methods.
  • To compare Between-Group Analysis and Discriminant Analysis (with dimension reduction).
  • To recommend an appropriate method based on dataset structure.

Main Methods:

  • Comparison of Between-Group Analysis and Discriminant Analysis (using Partial Least Squares or Principal Components Analysis for dimension reduction).

Related Experiment Videos

  • Geometric analysis to understand the relationship and data handling differences between methods.
  • Innovative simulation of datasets with controlled variance partitions (within-group and between-group) and use of public datasets.

    • Main Results:

    • Methods perform equally when clusters are clearly separated; both fail with overlapping clusters.
    • Discriminant Analysis is recommended for intermediate structures where projection handling is crucial.
    • A two-graph visualization tool is proposed for pre-analysis dataset structure examination.

    Conclusions:

    • Discriminant Analysis outperforms Between-Group Analysis by effectively utilizing dataset structure.
    • A priori knowledge of data structure guides the selection of appropriate analysis methods.
    • Simulated datasets are valuable for method assessment; caution against using unchallenging datasets like Golub for comparison.