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

Spectral analysis of two-signed microarray expression data.

Desmond J Higham1, Gabriela Kalna, J Keith Vass

  • 1Department of Mathematics, University of Strathclyde, Glasgow G1 1XH, UK. djh@maths.strath.ac.uk

Mathematical Medicine and Biology : a Journal of the IMA
|November 30, 2006
PubMed
Summary

This study introduces a spectral algorithm for analyzing gene expression data, revealing hidden patterns in DNA microarrays. The method effectively classifies tumors by distinguishing between gene up-regulation and down-regulation.

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

How adversarial attacks can disrupt seemingly stable accurate classifiers.

Neural networks : the official journal of the International Neural Network Society·2024
Same author

Estimating network dimension when the spectrum struggles.

Royal Society open science·2024
Same author

Connectivity of Random Geometric Hypergraphs.

Entropy (Basel, Switzerland)·2023
Same author

Generative hypergraph models and spectral embedding.

Scientific reports·2023
Same author

Epidemics on hypergraphs: spectral thresholds for extinction.

Proceedings. Mathematical, physical, and engineering sciences·2022
Same author

Directed network Laplacians and random graph models.

Royal Society open science·2021

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Genomics

Background:

  • Gene expression data from DNA microarrays captures simultaneous expression levels across samples.
  • Positive and negative weights signify gene up-regulation and down-regulation, respectively.
  • Unsupervised tumor classification is a key application for analyzing complex biological data.

Purpose of the Study:

  • To derive and present a spectral algorithm for clustering and reordering complementary DNA microarray expression data.
  • To provide theoretical support for the algorithm based on data structure hypotheses.
  • To demonstrate the algorithm's utility in unsupervised tumor classification.

Main Methods:

  • The algorithm is derived by solving a discrete optimization problem and then relaxing it to a continuous domain.

Related Experiment Videos

  • It involves imposing a random graph model on expression levels and clustering from a maximum likelihood perspective.
  • The method utilizes the dominant singular vectors to reveal 'checkerboard' sign patterns in the data.

    • Main Results:

    • The spectral algorithm successfully reveals inherent 'checkerboard' sign patterns in gene expression data.
    • The approach demonstrates tolerance to perturbations and identifies 'near-checkerboard' patterns.
    • Distinguishing between up-regulation and down-regulation provides insights not obtainable from absolute value data.

    Conclusions:

    • The developed spectral algorithm offers a novel approach to analyzing DNA microarray data.
    • The method's ability to differentiate gene regulation directions enhances biological structure discovery.
    • This technique shows promise for improving unsupervised tumor classification accuracy.