Validation and characterization of DNA microarray gene expression data distribution and associated moments.

Area of Science:

• Genomics
• Bioinformatics
• Statistical Biology

Background:

• DNA microarrays are crucial for understanding gene expression changes.
• Analyzing gene expression data requires robust distributional assumptions.
• Empirical validation of these assumptions is often overlooked.

Purpose of the Study:

• To empirically validate common distributional assumptions for gene expression data.
• To characterize observed gene expression distributions using statistical methods.
• To identify relationships between moments of gene expression distributions.

Main Methods:

• Utilized goodness-of-fit tests and moment-ratio diagrams.
• Analyzed data from the Gene Expression Omnibus (GEO) database.
• Included data from Affymetrix and Rosetta/Merck microarray platforms.

Main Results:

• Over 50% of probe sets on the Affymetrix platform rejected standard distributions.
• The Rosetta/Merck platform showed different rejection patterns, with 20% failing the logistic distribution test.
• Identified significant linear and quadratic relationships between distribution moments (mean, coefficient of variation, skewness, kurtosis).

Conclusions:

• Standard distributional assumptions are often invalid for gene expression data.
• Microarray platform and experimental conditions influence data distribution patterns.
• New statistical approaches are needed for accurate gene expression analysis.