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

Estimation of Sobol's Sensitivity Indices under Generalized Linear Models.

Rong Lu1, Danxin Wang2, Min Wang3

  • 1Bioinformatics Core Facility, Department of Clinical Sciences, University of Texas, Southwestern Medical Center, 5323 Harry Hines Blvd. Dallas, TX 75390.

Communications in Statistics: Theory and Methods
|September 22, 2018
PubMed
Summary

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

Pulmonary Rehabilitation with Airway Clearance Techniques Improves Mucus Hypersecretion and Clinical Outcomes in COPD: A Retrospective Study.

International journal of chronic obstructive pulmonary disease·2026
Same author

Knockdown of RNA editing proteins reshapes the HepaRG transcriptome and pharmacogene expression.

Scientific reports·2026
Same author

Separation-delayed skin flap for diabetic foot repair: a case report and review of the literature.

Journal of medical case reports·2026
Same author

Short-term Bayesian influenza forecasting in a hospital environment using a linear Kalman filter.

Journal of theoretical biology·2026
Same author

Infinitesimal homeostasis in mass-action systems.

Journal of mathematical biology·2026
Same author

Atorvastatin Protects Against Deleterious Carfilzomib-Induced Transcriptional Changes in Human Induced Pluripotent Stem Cell-Derived Cardiomyocytes.

International journal of molecular sciences·2026
This summary is machine-generated.

Sobol's sensitivity indices (SSIs) offer a computationally efficient method for variable selection in generalized linear models (GLMs). This approach identifies gene-gene interactions in microarray data, outperforming standard methods.

Area of Science:

  • Computational Statistics
  • Bioinformatics
  • Systems Biology

Background:

  • Generalized linear models (GLMs) are widely used for statistical analysis.
  • Variable selection and interaction identification are crucial in complex datasets like microarrays.
  • Existing methods for sensitivity analysis and variable selection can be computationally intensive.

Purpose of the Study:

  • To derive explicit formulas for Sobol's sensitivity indices (SSIs) within GLMs.
  • To establish SSIs as a tool for variable selection, particularly for polynomial regressions with identity links.
  • To apply SSI analysis for discovering gene-gene interactions in biological data.

Main Methods:

  • Derivation of analytical formulas for SSIs under GLMs with normal inputs.
Keywords:
Sobol’s indicescorrelated inputsgene-gene interactionsgeneralized linear modelsglobal sensitivity analysisvariable rankingvariable selection

Related Experiment Videos

  • Application of main-effect SSIs for variable selection in polynomial regression models.
  • Comparative analysis of SSI-based variable selection against the random forest algorithm.
  • Analysis of a public microarray dataset using SSI to identify gene-gene interactions.

    • Main Results:

    • Explicit formulas for SSIs are derived for GLMs with independent or multivariate normal inputs.
    • Main-effect SSIs demonstrate effectiveness for variable selection in GLMs with identity links.
    • SSI-based variable selection shows comparable results to random forest but with reduced computational cost.
    • Novel higher-order gene-gene interactions were identified in microarray data using SSI analysis.

    Conclusions:

    • SSIs provide a powerful and computationally efficient approach for variable selection in GLMs.
    • The derived SSI methods can uncover complex gene-gene interactions missed by traditional inference techniques.
    • An R package, SobolSensitivity, is available for implementing these SSI analysis functions.