randRotation

randRotation implements random rotation methods for performing exact testing of dependent test statistics with linear models following arbitrary batch effect correction on high-dimensional, normally distributed data with batch structure.

Key Features:

• Random Rotation Generation: Generates simulated datasets by random rotation to retain the dependence structure of genes adequately.
• Null Distribution Estimation: Estimates null distributions of dependent test statistics in data analysis pipelines where batch effect correction is followed by linear model analysis.
• Resampling-Based P-Value Calculation: Calculates resampling-based P-values and false-discovery rates following batch effect correction while maintaining the alpha level.
• Arbitrary Batch Effect Correction Compatibility: Works with arbitrary batch effect correction methods rather than requiring a single combined model.

Scientific Applications:

• Differential Gene Expression Analysis: Identifies differentially expressed genes with linear models following batch effect correction in microarray or RNA-seq data.
• Batch Effect Correction Validation: Addresses biased P-values or false discovery rates that arise when batch effect correction and hypothesis testing are done successively with separate models.
• High-Throughput Genomics Data Analysis: Analyzes data generated from sequencing, microarray and bead-chip technologies affected by batch effects.

Methodology:

The approach generates simulated datasets by random rotation to retain the dependence structure of genes, allowing estimation of null distributions of dependent test statistics for calculating resampling-based P-values and false-discovery rates following batch effect correction.