Fisher's combined p-value for detecting differentially expressed genes using Affymetrix expression arrays

Area of Science:

• Bioinformatics
• Genomics
• Statistical Genetics

Background:

• Traditional differential gene expression analysis relies on summary values, overlooking probe-level data.
• Emerging methods incorporate probe-level information for more nuanced analysis.
• Fisher's method offers a robust framework for combining evidence from multiple sources.

Purpose of the Study:

• To introduce and evaluate a novel method for differential gene expression analysis using probe-level data.
• To compare the performance of the proposed method against existing techniques.
• To highlight the utility of the method as a diagnostic tool for microarray data.

Main Methods:

• Development of a new approach combining p-values from probe-level significance tests using Fisher's method.
• Comparative analysis using spike-in datasets with known differentially expressed transcripts.
• Validation with biological array data and quantitative real-time PCR (qRT-PCR).

Main Results:

• The combined p-value method demonstrated favorable performance in terms of power and false discovery rates compared to competing methods on spike-in datasets.
• Probe-level testing methods, including the combined p-value approach, identified a substantial overlap in differentially expressed genes.
• The combined p-value method proved effective for diagnostic purposes in exploratory microarray data analysis.

Conclusions:

• The combined p-value method presents a promising alternative for differential gene expression testing.
• This method is particularly suitable for exploratory analysis and serves as a valuable diagnostic tool for microarray data.