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

Clustering threshold gradient descent regularization: with applications to microarray studies.

Shuangge Ma1, Jian Huang

  • 1Department of Epidemiology and Public Health, Yale University, New Haven, CT, USA. shuangge.ma@yale.edu

Bioinformatics (Oxford, England)
|December 22, 2006
PubMed
Summary

This study introduces a new method, clustering threshold gradient descent regularization (CTGDR), to identify key genes influencing clinical outcomes from microarray data by accounting for gene expression clusters. CTGDR improves feature selection for cancer classification and survival analysis.

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

Identification of sequence variants in genetic disease-causing genes using targeted next-generation sequencing.

PloS one·2012
Same author

Inhibition of matrine against gastric cancer cell line MNK45 growth and its anti-tumor mechanism.

Molecular biology reports·2011
Same author

Evolution of activation patterns during long-duration ventricular fibrillation in pigs.

American journal of physiology. Heart and circulatory physiology·2011
Same author

A new feruloyl amide derivative from the fruits of Tribulus terrestris.

Natural product research·2011
Same author

High-amylose rice improves indices of animal health in normal and diabetic rats.

Plant biotechnology journal·2011
Same author

The cross-validated AUC for MCP-logistic regression with high-dimensional data.

Statistical methods in medical research·2011

Area of Science:

  • Bioinformatics
  • Statistical Genetics
  • Computational Biology

Background:

  • Microarray studies aim to link gene expression to clinical outcomes like disease status and survival.
  • Gene expression data exhibit clustering, where genes within a cluster have correlated expressions and functions, but individual gene effects can differ.
  • Existing statistical models often lack sparsity or fail to adequately incorporate gene expression cluster structures.

Purpose of the Study:

  • To develop a sparse statistical model that accounts for gene expression cluster structures.
  • To identify genes significantly associated with clinical outcomes while considering their clustered nature.
  • To improve the accuracy of gene selection in microarray analyses.

Main Methods:

  • Genes are clustered using methods like K-means or hierarchical approaches, with the optimal number determined by the Gap statistic.

Related Experiment Videos

  • A novel Clustering Threshold Gradient Descent Regularization (CTGDR) method is proposed for simultaneous cluster and within-cluster gene selection.
  • The CTGDR method is applied to binary classification and censored survival analysis tasks.

    • Main Results:

    • The CTGDR method effectively performs feature selection at both the cluster and individual gene levels, outperforming standard methods.
    • The approach was successfully demonstrated on cancer classification and lymphoma patient survival studies using microarray expression data.
    • CTGDR's ability to leverage cluster structures enhances the identification of clinically relevant genes.

    Conclusions:

    • The CTGDR method provides a robust framework for sparse gene selection in microarray studies by integrating gene expression clustering.
    • This approach enhances the discovery of genes associated with clinical outcomes, offering improved insights for disease classification and prognosis.
    • The developed method offers a valuable tool for analyzing complex gene expression data in biomedical research.