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

A classification-based machine learning approach for the analysis of genome-wide expression data.

James Lyons-Weiler1, Satish Patel, Soumyaroop Bhattacharya

  • 1Department of Biological Sciences, University of Massachusetts, Lowell, Massachusetts 01854, USA. lyonsweilerj@msx.upmc.edu

Genome Research
|March 6, 2003
PubMed
Summary

The Maximum Difference Subset (MDSS) algorithm enhances gene expression analysis by integrating prediction accuracy and statistical significance. This approach identifies reliable biomarkers for predicting chemotherapy outcomes in acute myeloid leukemia.

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

Triggering of viral and bacterial respiratory infection hospitalizations by traffic pollution exposure in a cohort of hospitalized adults.

PloS one·2026
Same author

Evidence-Based Use of Undenatured Type II Collagen in the Management of Osteoarthritis in the Indian Subcontinent and Southeast Asia: A Narrative Review and an Expert Opinion.

Cureus·2026
Same author

Discriminating Bacterial from Nonbacterial Lower Respiratory Tract Infection within Clinical Subgroups of Hospitalized Adults.

The Journal of infectious diseases·2026
Same author

Type 2 Autoimmune Pancreatitis in a Young Adult Male: A Diagnostic Challenge.

Cureus·2026
Same author

Integrated analysis of insulin resistance reveals metabolic remodeling following diet switch-triggered calorie reduction.

Science advances·2026
Same author

Case Report: Post-stroke hemorrhagic infarction in a status epilepticus Beagle dog.

Frontiers in veterinary science·2026

Area of Science:

  • Genomics
  • Bioinformatics
  • Computational Biology

Background:

  • Global gene expression analysis involves class discovery, prediction, and biomarker identification.
  • Clinical applications require marker genes for accurate disease subclass prediction using microarray data.
  • Existing methods include clustering, statistical tests, and machine learning.

Purpose of the Study:

  • To introduce the Maximum Difference Subset (MDSS) algorithm for gene expression data analysis.
  • To integrate classification, statistics, and machine learning for a coherent analytical framework.
  • To improve the accuracy and external validity of predictive gene sets.

Main Methods:

  • The MDSS algorithm combines classification algorithms, classical statistics, and machine learning.

Related Experiment Videos

  • It integrates prediction accuracy to determine the critical threshold of statistical significance (P-value).
  • A jackknife step is employed to reduce false positives and enhance external validity by removing genes with low predictive utility.

    • Main Results:

    • The MDSS algorithm learns the significance level relevant to the dataset, minimizing reliance on arbitrary thresholds.
    • It demonstrates high external validity, with predictions less dependent on study design.
    • Unlike other methods, MDSS identified biomarkers for predicting anthracycline-cytarabine chemotherapy outcomes in acute myeloid leukemia.

    Conclusions:

    • The MDSS approach offers a robust framework for gene expression analysis, combining statistical significance and predictive utility.
    • It enhances the reliability and external validity of biomarker discovery.
    • MDSS can be applied with various test and classifier operator pairs for diverse applications.