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 protocol for building and evaluating predictors of disease state based on microarray data.

Lodewyk F A Wessels1, Marcel J T Reinders, Augustinus A M Hart

  • 1Department of Mediamatics, Faculty of Electrical Engineering, Mathematics and Computer Science, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands. l.f.a.wessels@ewi.tudelft.nl

Bioinformatics (Oxford, England)
|April 9, 2005
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

Tremelimumab with or without durvalumab in combination with paclitaxel in metastatic urothelial cancer: phase I/II ICRA trial.

Nature communications·2026
Same author

Enabling DCIS subtyping: leveraging foundation models for robust grading and molecular biomarker scoring.

NPJ breast cancer·2026
Same author

From prediction to precision: Biomarker discovery and predictive modeling for personalized immune checkpoint blockade therapy.

European journal of cancer (Oxford, England : 1990)·2026
Same author

Associations between metal mixture exposure and emotional distress symptoms among pregnant women: A prospective study.

Journal of trace elements in medicine and biology : organ of the Society for Minerals and Trace Elements (GMS)·2026
Same author

Cross-cultural adaptation and psychometric evaluation of the Chinese adapted and revised food allergy self-efficacy scale for parents.

Frontiers in psychology·2026
Same author

Cell-State-Specific Drug Responses are Associated With Differences in Signaling Network Wiring.

Molecular & cellular proteomics : MCP·2026
Same journal

Probabilistic RNA designability via interpretable ensemble approximation and dynamic decomposition.

Bioinformatics (Oxford, England)·2026
Same journal

Quantifying domain-specific relevance of computational biology Wikipedia articles using TF-IDF and cosine similarity.

Bioinformatics (Oxford, England)·2026
Same journal

GATSBI: improving context-aware protein embeddings through biologically motivated data splits.

Bioinformatics (Oxford, England)·2026
Same journal

BiMba: using Vision Mamba to predict protein sites that bind other proteins.

Bioinformatics (Oxford, England)·2026
Same journal

ProMeta: a meta-learning framework for robust disease diagnosis and prediction from plasma proteomics.

Bioinformatics (Oxford, England)·2026
Same journal

Is a Win-Win possible? Achieving pareto-optimal privacy-utility balance in fine-tuned genome language model embeddings against embedding reconstruction attacks.

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

Developing robust cancer outcome prediction models requires objective evaluation. This study introduces a validated protocol, demonstrating simple methods like shrunken centroids outperform complex ones for accurate gene expression analysis.

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Cancer Genomics

Background:

  • Microarray gene expression data are vital for identifying cancer biomarkers.
  • Numerous computational methods exist for predictor construction, but lack objective evaluation standards.
  • No universally accepted computational approach ensures reliable outcome prediction.

Purpose of the Study:

  • To introduce a principled training and validation protocol for evaluating cancer outcome prediction methodologies.
  • To objectively assess different reporter selection strategies and predictor choices using gene expression data.

Main Methods:

  • Development and application of a standardized training and validation protocol.
  • Evaluation of reporter selection strategies (e.g., forward filtering, shrunken centroids) and predictors (e.g., nearest mean classifier, partial least squares).

Related Experiment Videos

  • Testing on six diverse gene expression datasets to assess performance across varying difficulties.

    • Main Results:

    • The proposed protocol enables objective evaluation of predictor construction methodologies.
    • Simple reporter selection strategies (forward filtering, shrunken centroids) were effective, outperforming partial least squares on four of six datasets.
    • Simpler classifiers, such as the nearest mean classifier, demonstrated superior performance compared to more complex alternatives.

    Conclusions:

    • The developed protocol offers a robust framework for evaluating computational approaches in cancer outcome prediction.
    • Objective comparison of different methods is now feasible, facilitating the advancement of reliable predictive models.
    • Simple, validated methods can achieve high performance in predicting cancer outcomes from gene expression data.