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

Accounting for probe-level noise in principal component analysis of microarray data.

Guido Sanguinetti1, Marta Milo, Magnus Rattray

  • 1Department of Computer Science, Regent Court 211 Portobello Road, Sheffield S1 4DP, UK.

Bioinformatics (Oxford, England)
|August 11, 2005
PubMed
Summary

This study introduces a novel model-based Principal Component Analysis (PCA) accounting for gene-specific variances in microarray data. This approach enhances data denoising and automatically determines the optimal number of principal components.

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

Multimodal spatial omics: From data acquisition to computational integration.

Patterns (New York, N.Y.)·2026
Same author

A cell atlas of the developing human outflow tract of the heart and its adult aortic valve derivatives.

eLife·2026
Same author

DeepSynBa: actionable drug combination prediction with complete dose-response profiles.

Bioinformatics (Oxford, England)·2026
Same author

Scalable, fast and accurate differential gene expression testing from millions of cells of multiple patients.

Nature communications·2026
Same author

metaAPA: a tool for integration of PolyA site predictions from single-cell and spatial transcriptomics.

Bioinformatics advances·2026
Same author

CRAK-Velo: chromatin accessibility kinetics integration improves RNA velocity estimation.

Genome biology·2026

Area of Science:

  • Bioinformatics
  • Computational Biology
  • Statistical Genetics

Background:

  • Principal Component Analysis (PCA) is a common dimensionality reduction technique for high-dimensional data.
  • Standard PCA does not account for data point error measures beyond spherical noise, limiting its use with variable biological data like microarrays.
  • Credibility intervals from probe-level analysis of microarray experiments can be gene and experiment specific.

Purpose of the Study:

  • To develop a model-based Principal Component Analysis (PCA) that incorporates variances associated with each gene and experiment.
  • To improve the analysis of high-dimensional biological datasets, particularly microarray data.
  • To provide a more accurate and robust dimensionality reduction method for biological data.

Main Methods:

Related Experiment Videos

  • A novel model-based approach to Principal Component Analysis (PCA) was proposed.
  • An efficient Expectation-Maximization (EM) algorithm was developed for parameter estimation.
  • The method was applied to microarray datasets to assess its performance.

Main Results:

  • The proposed model-based PCA significantly outperforms standard PCA.
  • The method effectively 'denoises' microarray datasets, improving expression profiles.
  • Tighter clustering across expression profiles was observed, indicating enhanced data organization.
  • The probabilistic nature of the model automatically determines the correct number of principal components.

Conclusions:

  • The developed model-based PCA offers a computationally efficient and statistically robust alternative to standard PCA for high-dimensional biological data.
  • This approach enhances the reliability of dimensionality reduction in microarray analysis.
  • The method facilitates improved data interpretation and biological discovery through better denoising and clustering.