Graphical Exploration of Gene Expression Data: A Comparative Study of Three Multivariate Methods

• Published in: Biometrics Vol. 59; no. 4; pp. 1131 - 1139
• Main Authors: Wouters, Luc, Göhlmann, Hinrich W., Bijnens, Luc, Kass, Stefan U., Molenberghs, Geert, Lewi, Paul J.
• Format: Journal Article
• Language: English
• Published: 350 Main Street , Malden , MA 02148 , U.S.A , and P.O. Box 1354, 9600 Garsington Road , Oxford OX4 2DQ , U.K Blackwell Publishing 01.12.2003; International Biometric Society
• Subjects: Bioinformatics; Biometry - methods; Biplot; Consultant's Forum; Correspondence factor analysis; Data mining; Data visualization; Datasets; Gene Expression; Gene expression data; Genes; Genetic mapping; Leukemia; Microarray data; Models, Genetic; Models, Statistical; Multivariate Analysis; Multivariate exploratory data analysis; Oligonucleotide Array Sequence Analysis - methods; Principal component analysis; Principal components analysis; Reproducibility of Results; Spectral map analysis; Spectroscopic analysis; Statistical variance; T lymphocytes; Term weighting
• ISSN: 0006-341X; 1541-0420
• DOI: 10.1111/j.0006-341X.2003.00130.x

This article describes three multivariate projection methods and compares them for their ability to identify clusters of biological samples and genes using real-life data on gene expression levels of leukemia patients. It is shown that principal component analysis (PCA) has the disadvantage that the resulting principal factors are not very informative, while correspondence factor analysis (CFA) has difficulties interpreting distances between objects. Spectral map analysis (SMA) is introduced as an alternative approach to the analysis of microarray data. Weighted SMA outperforms PCA, and is at least as powerful as CFA, in finding clusters in the samples, as well as identifying genes related to these clusters. SMA addresses the problem of data analysis in microarray experiments in a more appropriate manner than CFA, and allows more flexible weighting to the genes and samples. Proper weighting is important, since it enables less reliable data to be down-weighted and more reliable information to be emphasized.