Application of One Sided t-tests and a Generalized Experiment Wise Error Rate to High-Density Oligonucleotide...

Area of Science:

• Molecular Biology
• Genetics
• Bioinformatics

Background:

• Analyzing microarray data to find differentially expressed genes presents a significant challenge.
• Standard statistical methods may lack the power to detect subtle expression changes.

Purpose of the Study:

• To evaluate statistical methods including ANOVA, one-sided t-tests, and a generalized family-wise error rate (GFWER) for microarray data analysis.
• To identify genes affected by the PICKLE chromatin remodeling factor in Arabidopsis thaliana.

Main Methods:

• Applied ANOVA, one-sided t-tests, natural log transformation, and GFWER methodology to analyze Affymetrix GeneChip data.
• Utilized a generalized experiment-wise error rate (GFWER(k)) to control false rejections.
• Performed quantitative real-time PCR (qRT-PCR) for validation.

Main Results:

• The GFWER(k) method demonstrated simplicity and increased statistical power.
• Using a one-sided t-test with GFWER(2)=0.05 identified 43 genes with PICKLE-dependent expression.
• qRT-PCR confirmed PICKLE-dependent expression in 83.7% (36 out of 43) of the identified genes.

Conclusions:

• The GFWER(k) approach, particularly with k=2 or 3, is effective for identifying differentially expressed genes in microarray experiments.
• This methodology enhances the power to detect gene expression changes, as validated by qRT-PCR.