Testing for treatment effects on gene ontology

Area of Science:

• Genomics
• Bioinformatics
• Systems Biology

Background:

• Gene expression studies using DNA arrays benefit from analyzing functional gene groups.
• Gene Ontology (GO) terms provide a framework for categorizing genes by function.
• Interpreting treatment effects is enhanced by assessing significance at the pathway level.

Purpose of the Study:

• To develop a modified meta-analysis method for assessing treatment effects on functionally defined gene groups.
• To incorporate gene correlation structures for increased significance and specificity in pathway analysis.
• To validate the method's performance in identifying altered pathways using microarray data.

Main Methods:

• A modified meta-analysis approach was developed to combine p-values from gene expression data.
• The method accounts for the correlation structure among genes within functional groups.
• P-values are calculated using a Monte Carlo method for hypothesis testing with two-directional gene expression changes.

Main Results:

• The developed analytic method demonstrated improved specificity in selecting altered pathways.
• Incorporation of inter-gene correlation structure within pathways enhanced the detection of significant effects.
• The method provided reasonable test results for complex microarray designs and small sample sizes.

Conclusions:

• The modified meta-analysis method is a practical approach for measuring treatment effects on GO groups.
• The technique enhances the interpretation of gene expression data by focusing on functional pathways.
• This method offers improved statistical power and specificity for analyzing microarray experiments.