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

Modeling gene expression measurement error: a quasi-likelihood approach.

Korbinian Strimmer1

  • 1Department of Statistics, University of Munich, Ludwigstrasse 33, D-80539 Munich, Germany. strimmer@stat.uni-muenchen.de

BMC Bioinformatics
|March 28, 2003
PubMed
Summary

A new semi-parametric model using extended quasi-likelihood offers a flexible approach to gene expression data analysis. This method improves data fitting and enhances the power of differential expression tests without strict distributional assumptions.

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

Quantification of protein abundance and interaction defines a mechanism for operation of the circadian clock.

eLife·2022
Same author

A whitening approach to probabilistic canonical correlation analysis for omics data integration.

BMC bioinformatics·2019
Same author

Differential protein expression and peak selection in mass spectrometry data by binary discriminant analysis.

Bioinformatics (Oxford, England)·2015
Same author

A simple data-adaptive probabilistic variant calling model.

Algorithms for molecular biology : AMB·2015
Same author

A novel algorithm for simultaneous SNP selection in high-dimensional genome-wide association studies.

BMC bioinformatics·2012
Same author

Signal identification for rare and weak features: higher criticism or false discovery rates?

Biostatistics (Oxford, England)·2012

Area of Science:

  • Genomics
  • Statistical Bioinformatics
  • Computational Biology

Background:

  • Accurate error modeling is crucial for analyzing gene expression data from microarrays.
  • Current methods often assume simple parametric models or empirical distributions, which may be suboptimal.
  • These approaches can ignore structural information or risk misspecification, and scale choice (linear vs. log) is problematic.

Purpose of the Study:

  • To introduce a novel semi-parametric model for gene expression measurement errors.
  • To develop a statistical framework that requires only partial knowledge of the underlying error distribution.
  • To offer a versatile method for analyzing gene expression data on any scale and incorporating systematic effects.

Main Methods:

  • Developed a semi-parametric model based on the extended quasi-likelihood function.

Related Experiment Videos

  • Utilized the postulated variance structure of gene expression data (e.g., quadratic) to construct the likelihood.
  • Applied the framework for parameter estimation, confidence interval calculation, and regression analysis of systematic effects.

    • Main Results:

    • The extended quasi-likelihood provides a robust method for inference, similar to a proper likelihood.
    • The model allows for estimation of calibration and variance parameters with approximate confidence intervals.
    • Analysis can be performed on both linear and transformed scales, accommodating systematic effects like array or dye biases.

    Conclusions:

    • The quasi-likelihood framework offers a simple, versatile, and assumption-light approach for gene expression data analysis.
    • Demonstrated superior data fit compared to existing models on simulated and real datasets.
    • Showcased improved statistical power for identifying differentially expressed genes in practical applications.