maanova

maanova performs ANOVA-based statistical analysis of two-color cDNA microarray experiments to estimate differential gene expression and model variance across complex and multi-factorial experimental designs.

Key Features:

• Implementation: Implemented in R and Matlab.
• Data quality assessment: Provides functions for assessing two-color cDNA microarray data quality.
• Normalization and transformation: Supports data normalization and transformation prior to analysis.
• ANOVA-based linear models: Estimates relative gene expression using ANOVA-based linear models for complex experimental designs.
• Hypothesis testing: Performs formal hypothesis testing for differential expression.
• Permutation- and shrinkage-based tests: Implements permutation-based and shrinkage-based statistical tests to enable control of false discovery rate across thousands of genes.
• Model evaluation and diagnostics: Offers tools for model evaluation and diagnostic assessment.
• Graphical summaries: Produces graphical summaries of expression patterns.
• Cluster analysis with bootstrapping: Supports robust cluster analysis via bootstrapping.
• Extensibility: Functions can be integrated with other statistical and visualization tools.

Scientific Applications:

• Two-color cDNA microarray analysis: Analysis of gene expression data generated by two-color cDNA microarray experiments.
• Differential expression: Identification of genes differentially expressed across experimental conditions.
• Complex and multi-factorial designs: Modeling variance and effects in complex, multi-factorial experimental designs.
• False discovery rate control: Controlling FDR in high-dimensional gene expression studies using permutation and shrinkage methods.
• Cluster analysis and pattern discovery: Discovery of expression patterns and clusters with bootstrapping for robustness.
• Model diagnostics and validation: Evaluation and diagnostic assessment of fitted statistical models.

Methodology:

Data quality assessment, normalization and transformation, ANOVA-based linear modeling to estimate relative gene expression, formal hypothesis testing including permutation-based and shrinkage-based tests for FDR control, model evaluation and diagnostics, graphical summaries, and bootstrapped cluster analysis.