Tools for Connectivity Map-like analyses
Description
The gCMAP package provides a toolkit for comparing
differential gene expression profiles through gene set
enrichment analysis. Starting from normalized microarray or
RNA-seq gene expression values (stored in lists of
ExpressionSet and CountDataSet objects) the package performs
differential expression analysis using the limma or DESeq
packages. Supplying a simple list of gene identifiers, global
differential expression profiles or data from complete
experiments as input, users can use a unified set of several
well-known gene set enrichment analysis methods to retrieve
experiments with similar changes in gene expression. To take
into account the directionality of gene expression changes,
gCMAPQuery introduces the SignedGeneSet class, directly
extending GeneSet from the GSEABase package. To increase
performance of large queries, multiple gene sets are stored as
sparse incidence matrices within CMAPCollection eSets. gCMAP
offers implementations of 1. Fisher's exact test (Fisher, J R
Stat Soc, 1922) 2. The "connectivity map" method (Lamb et al,
Science, 2006) 3. Parametric and non-parametric t-statistic
summaries (Jiang & Gentleman, Bioinformatics, 2007) and 4.
Wilcoxon / Mann-Whitney rank sum statistics (Wilcoxon,
Biometrics Bulletin, 1945) as well as wrappers for the 5.
camera (Wu & Smyth, Nucleic Acid Res, 2012) 6. mroast and romer
(Wu et al, Bioinformatics, 2010) functions from the limma
package and 7. wraps the gsea method from the mgsa package
(Bauer et al, NAR, 2010). All methods return CMAPResult
objects, an S4 class inheriting from AnnotatedDataFrame,
containing enrichment statistics as well as annotation data and
providing simple high-level summary plots.