Skip to main navigation Skip to search Skip to main content

Gaussian mixture decomposition of time-course DNA microarray data

  • Rice University
  • Maria Sklodowska-Curie Institute of Oncology

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

5 Citations (Scopus)

Abstract

In this chapter we present the decomposition approach to the analysis of large gene expression profile data sets. We address the problem of analysis of transient time-course data of expression profiles. We accept the assumption that co-expression of genes can be related to their belonging to the same Gaussian component. We assume that parameters of Gaussian components, means and variances, can differ between time instants. However, the gene composition of components is unchanged between time instants. For such problem formulations we derive the appropriate version of expectation-maximization algorithm recursions for the estimation of model parameters.We apply the derived method to the data on gene expression profiles of human K562 erythroleukemic cells and we discuss the obtained gene clustering.

Original languageEnglish
Title of host publicationModeling and Simulation in Science, Engineering and Technology
PublisherSpringer Basel
Pages351-359
Number of pages9
DOIs
Publication statusPublished - 2007

Publication series

NameModeling and Simulation in Science, Engineering and Technology
Volume38
ISSN (Print)2164-3679
ISSN (Electronic)2164-3725

Keywords

  • EM method
  • Gaussian components
  • Gene expression profiles
  • maximum likelihood

ASJC Scopus subject areas

  • Modeling and Simulation
  • General Engineering
  • Fluid Flow and Transfer Processes
  • Computational Mathematics

Fingerprint

Dive into the research topics of 'Gaussian mixture decomposition of time-course DNA microarray data'. Together they form a unique fingerprint.

Cite this