|
Activity Number:
|
61
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 3, 2003 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #300209 |
|
Title:
|
Spectral Map Analysis--A Method To Analyze Gene Expression Data
|
|
Author(s):
|
Luc Bijnens*+ and Paul Lewi and Hinrich Goehlmann and Geert Molenberghs and Luc Wouters
|
|
Companies:
|
Janssen Pharmaceutica and Janssen Pharmaceutica and Johnson & Johnson Pharmaceutical Research and Development, LLC and Limburgs University Centrum and Barrier Therapeutics NV
|
|
Address:
|
Turnhoutseweg 30, Beerse, , B-2340, Belgium
|
|
Keywords:
|
gene expression data ; microarrays ; correspondence factor analysis ; principle component analysis ; spectral map analysis
|
|
Abstract:
|
The simultaneous measurement of the expression level of thousands of genes presents a real challenge to existing statistical software tools because of the complexity and size of the datasets. In this study, three multivariate data analysis methods: principal component analysis (PCA), correspondence factor analysis (CFA), and spectral map analysis (SMA) are compared for their ability to identify clusters of biological samples and genes using data on gene expression levels of leukemia patients (Golub et al. 1999). PCA has the disadvantage that the resulting principal factors are not very informative regarding differential gene expression, while CFA is sensitive to single large values and has difficulties regarding interpretation of the distances between objects. We present SM, an alternative method developed by Lewi (1976). The importance of weighting for the level of gene expression is demonstrated. Proper weighting allows less reliable data to be downweighted and more reliable information to be emphasized. It is shown that weighted SMA outperforms PCA and CFA in finding clusters in the biological samples and identifying genes related to these clusters.
|
