Factor analysis of cluster-specific gene expression levels from cDNA microarrays

Leif E. Peterson

Research output: Contribution to journalArticlepeer-review

31 Scopus citations


The ever-increasing use of cDNA microarrays in medical research will require the development of new algorithms designed specifically for desktop analysis of potentially large genetic data sets. This paper describes the CLUSFAVOR algorithm (CLUSter and Factor Analysis Using Varimax Orthogonal Rotation) for performing cluster and factor analysis of gene expression data obtained from cDNA microarrays. A unique feature of the CLUSFAVOR algorithm is that a user can perform cluster analysis, view dendograms, and run factor analysis on cluster-specific genes selected interactively within a single executable program. CLUSFAVOR also performs a varimax orthogonal rotation on the extracted factors to increase parsimony in the loadings, revealing unique expression profiles for genes and expressed sequence tags (ESTs) for which pathway and function information is unknown. Microarray data used by the program must be stored and input from a disk file. Optional output contains matrices for the input data, standardized data, distance matrices, factor loadings, eigenvalues, eigenvectors, and percentage of total variation for genes within a cluster. Color cluster image displays containing gene expression levels, dendograms for arrays and genes, and factor loadings are displayed for the entire group of genes and arrays analyzed as well as selected genes within a cluster. Cluster images are also exported to JPG and linked to HTML files for viewing.

Original languageEnglish (US)
Pages (from-to)179-188
Number of pages10
JournalComputer Methods and Programs in Biomedicine
Issue number3
StatePublished - Nov 2002


  • cDNA microarrays
  • Cluster analysis
  • Eigenanalysis
  • Eigenvalues
  • Eigenvectors
  • Factor analysis
  • Gene expression
  • Hierarchical agglomeration

ASJC Scopus subject areas

  • Software


Dive into the research topics of 'Factor analysis of cluster-specific gene expression levels from cDNA microarrays'. Together they form a unique fingerprint.

Cite this