gaga

gaga implements hierarchical gamma–gamma models for differential expression analysis of microarray and high-throughput data to improve sensitivity and accuracy with small to moderate sample sizes.

Key Features:

• Differential Expression Analysis: Utilizes hierarchical gamma–gamma models to detect differentially expressed genes with increased sensitivity and accuracy for microarray and high-throughput data.
• Supervised Gene Clustering and Classification: Provides supervised clustering and classification of gene expression data.
• Sequential Sample Size Calculations: Performs sequential sample size calculations using GaGa and LNNGV models, with LNNGV available from the EBarrays package.
• Model Extensions: Offers a simple extension to the gamma–gamma model and a more complex extension using a mixture of gamma distributions to improve model fit.
• Computational Efficiency: Derives approximations to reduce computational costs while maintaining model performance.

Scientific Applications:

• Differential expression in microarray and high-throughput studies: Detects differentially expressed genes, particularly in studies with small to moderate sample sizes.
• Experimental design and sample size estimation: Supports sequential sample size calculations for planning experiments using GaGa and LNNGV models.
• Gene clustering and classification: Applies supervised clustering and classification for organization and interpretation of gene expression patterns.

Methodology:

Implements the gamma–gamma hierarchical model (Kendziorski et al., Newton et al.) with a simple extension and a mixture-of-gamma-distributions extension, derives computational approximations, and applies GaGa and LNNGV models (LNNGV from EBarrays) for sequential sample size calculations.