Changes in version 1.19.4:

    o   CHANGE: As the bigmemory package is not available for Windows, only
	 unix_type OS is supported.

Changes in version 1.19.1:

    o   BUGFIX: Updated NAMESPACE file to conform with R CMD check.

    o   BUGFIX: reactome2cmap function is available again.

    o   NEW: Added citation information.

Changes in version 1.17.3:

    o   BUGFIX: Simpified the mergeCMAPs function.

    o   BUGFIX: Added additional imports to NAMESPACE, as recommended by R
	 CMD check.

Changes in version 1.17.2:

    o   BUGFIX: Added import of the Matrix package to NAMESPACE file.

Changes in version 1.17.0:

    o   BUGFIX: Added camera's newly introduced inter.gene.cor parameter in
	 camera_score-methods.

Changes in version 1.11.3:

    o   BUGFIX: The wiki2cmap function now excludes empty files downloaded
	 from wikipathways from parsing.

    o   BUGFIX: The go2cmap function is now exported.

    o   BUGFIX: The reactom2cmap function is currently defunct, because the
	 reactome.db package is broken.

Changes in version 1.11.2:

    o   CHANGE: Fixed typo in mroast_score method.

Changes in version 1.11.1:

    o   BUGFIX: Fixed the romer_score method to accomodate the change in
	 argument order in limma's romer function.

Changes in version 1.11.0:

    o   CHANGE: Version bump for BioC update.

Changes in version 1.10.0:

    o   CHANGE: Version bump for BioC update.

Changes in version 1.9.2:

    o   BUGFIX: Updated geneIndex method to accept both symbols2indices and
	 id2indices functions from limma, depending on the limma version.

Changes in version 1.8.1:

    o   BUGFIX: Updated extraction of big.max objects while maintaining
	 row.names.

Changes in version 1.7.2:

    o   CHANGE: Helper functions KEGG2cmap, reactome2cmap, wiki2cmap and
	 go2cmap have two additional parameters, min.size and max.size,
	 restricting the output to gene sets containing members within the
	 specified range.

    o   CHANGE: Method camera_score now outputs correlations in column
	 'effect'.

    o   BUGFIX: Correct numbers of annotated genes are now returned for each
	 set.

Changes in version 1.7.1:

    o   CHANGE: Vignettes are now stored in vignettes directory.

Changes in version 1.5.4:

    o   NEW: The bigmemory and bigmemoryExtras packages are not optional,
	 enabling use of gCMAP on Windows

Changes in version 1.5.3:

    o   NEW: Updated titles of vignettes

Changes in version 1.5.2:

    o   BUGFIX: Updated eSet construction in mapNmerge function

Changes in version 1.5.1:

    o   BUGFIX: fixed incorrect varMetaData element definition in
	 .process_counts

Changes in version 1.3.5:

    o   NEW: DESeq package is now listed as 'Suggested' instead of 'Depends'

Changes in version 1.3.4:

    o   NEW: Removed lattice and reshape dependencies.

    o   NEW: Removed cmapQAPlot function.

    o   NEW: Added generate_gCMAP_NChannelSet's 'report.center' parameter.

Changes in version 1.3.3:

    o   NEW: Experimented with cmapQAPlot function (removed again in version
	 1.3.4).

Changes in version 1.3.2:

    o   NEW: mgsa_score method wrapping the model-based gene set analysis
	 approach available in the mgsa package.

Changes in version 1.3.1:

    o   NEW: KEGG2cmap, reactome2cmap, wiki2cmap and go2cmap to functions to
	 generate CMAPCollections from public annotation databases

    o   NEW: upIds and downIds methods for CMAPCollection objects.

Changes in version 1.3.0:

    o   NEW: fisher_score method now accepts NChannelSets as 'sets' and
	 thresholds a channel of choice on the fly.

    o   BUGFIX: induceCMAPCollection method was updated to fix a bug that
	 prevented processing of certain single-sample eSets.

    o   NEW: parameters 'center.z' and 'center.log_fc' for the
	 generate_gCMAP_NChannelSet function, centering each column of the
	 score matrix on the column mean, median or density peak.
	 (Internally, the NEW: 'center_eSet' function is called.)

    o   NEW: function 'center_eSet' to center columns of eSet channels on
	 either their kernel density peak, their mean or their median.

    o   NEW: function mergeCMAPs to combined two eSets that share class,
	 varLabels, annotation and assayDataElementNames.

    o   BUGFIX: Fixed missing gene names in featureScores method with
	 CMAPCollections queries containing single-gene sets.

    o   NEW: splitPerturbations function to split ExpressionSet (e.g.
	 downloaded from ArrayExpress) by annotations provided in the
	 phenoData slot.

    o   NEW: eset_instances function to generate lists eSets from larger
	 datasets by specifying an instance.matrix.

    o   NEW: setSizes methods for CMAPCollections returns a data.frame with
	 the number of members for each gene set stored in the Collection.

    o   NEW: fisher_score method now reports z-scores in the effect column
	 instead of log-odds-ratio. The latter are returned in the NEW: ,
	 separate LOR column.

    o   NEW: featureScores method now return list of matched identfiers when
	 dispatched on two CMAPCollection.

    o   NEW: Added 'density.only', 'level.only' and 'strip.only' parameters
	 to the CMAPResults plot method to selectively return only one plot
	 component.

    o   IMPROVED Reduced memory footprint of gsealm_jg_score and
	 induceCMAPCollection methods.

    o   CHANGED: By default, 'removeShift' option for gsealm_jg_score and
	 gsealm_score methods is now set to FALSE.

    o   BUGFIX: Corrected sign of effect size for fisher_score for perfect
	 overlap between set and query.

    o   IMPROVED: Improved plot method for CMAPResults objects, lattice
	 package is no longer required.

Changes in version 1.1.4:

    o   BUGFIX: The gealm_score method is not defined for queries with
	 missing values and will now return the corresponding error message.

    o   BUGFIX: GeneSetCollections containing only one set are processed
	 correctly by gsealm_score.

Changes in version 1.1.3:

    o   BUGFIX: Corrected sign of effect size for fisher_score for perfect
	 overlap between set and query.

Changes in version 1.1.2:

    o   BUGFIX: Fixed log-odds calculation in fisher_score: non-overlapping
	 sets now get a log odds ratio of 0

Changes in version 1.1.1:

    o   IMPROVED: Reduced memory footprint of induceCMAPCollection method
	 for bigmemory-backed eSet objects.

Changes in version 1.1.0:

    o   NEW: Version number was increased for the Bioconductor release.

Changes in version 0.99.6:

    o   IMPROVED: Updated mapIdentifiers method for SignedGeneSet class now
	 handles more different annotation types.

Changes in version 0.99.5:

    o   NEW: The BigMatrix package changed its name to bigmemoryExtras.
	 Currently, bigmemoryExtras only supports Mac and Unix platforms,
	 thus gCMAP does not support Windows, either.

Changes in version 0.99.1:

    o   NEW: First release of the gCMAP package to Bioconductor.