Mulcom

Mulcom implements a modified Dunnett's t-test as an R-Bioconductor package to identify differentially expressed genes in microarray experiments comparing multiple test groups against a common reference.

Key Features:

• Enhanced Statistical Methodology: Applies a modified Dunnett's t-test that aggregates within-group variance from all experimental groups rather than estimating standard error from pairwise comparisons.
• False Discovery Rate (FDR) Optimization: Uses permutation-based estimation of false discovery rate (FDR) to control false positives and maximize significant gene discovery at a specified FDR level.
• Minimal Fold-Change Threshold: Supports an optional minimal fold-change threshold parameter (m) to filter results for biological relevance.
• Cross-Platform Consistency and Functional Enrichment: Demonstrated higher concordance in identifying significant genes across microarray platforms (39% vs 26% or 15%) and greater enrichment in functionally relevant gene categories.

Scientific Applications:

• Multiple-group vs. baseline comparisons: Identifying differentially expressed genes when several experimental conditions are compared to a single reference.
• Time-course studies: Detecting genes with differential expression across time points relative to baseline.
• Drug treatment effects: Characterizing gene expression changes induced by treatments compared to control.
• Disease progression analyses: Comparing stages or conditions to a reference to find progression-associated expression changes.
• Cross-platform microarray analysis: Obtaining consistent differential expression results and functional enrichment across different microarray platforms.

Methodology:

Adapts Dunnett's t-test to compare each test group against the reference while pooling within-group variance across all groups; performs permutation-based FDR estimation; and applies an optional minimal fold-change threshold (m).