Analysis of Correlated Gene Expression Data on Ordered Categories.

Area of Science:

• Bioinformatics
• Statistical Genetics
• Computational Biology

Background:

• Analyzing longitudinal microarray gene expression data presents challenges due to complex correlation structures.
• Existing methods may not adequately address both intra-chip gene correlations and temporal correlations within subjects.
• Order-restricted inference is crucial for comparing gene expression across ordered experimental conditions.

Purpose of the Study:

• To develop a non-parametric bootstrap-based methodology for analyzing repeated measures/longitudinal microarray gene expression data.
• To implement a procedure that accounts for both intra-chip and temporal correlations.
• To ensure the method controls the false discovery rate (FDR) for reliable significance testing.

Main Methods:

• A bootstrap-based non-parametric procedure utilizing order-restricted inference.
• Bootstrapping residuals to derive the null distribution for significance testing.
• Incorporation of the adaptive bootstrap methodology (Guo and Peddada, 2008) for computational efficiency.

Main Results:

• The proposed methodology effectively analyzes longitudinal gene expression data over ordered categories.
• The procedure successfully addresses intra-chip and temporal correlation structures.
• The adaptive bootstrap implementation ensures control of the false discovery rate (FDR).

Conclusions:

• The introduced bootstrap methodology provides a robust framework for analyzing complex longitudinal gene expression data.
• This approach enhances the reliability of identifying significant gene expression changes in ordered experimental settings.
• The method offers a computationally efficient and statistically sound tool for genomic research.