Changes in version 1.28.0                        

    o   Minor optimization to updateObject() to return early if the
	object is more recent than the package version.

                       Changes in version 1.14.0                        

    o   Added the unsplitAltExps() function to reverse the effect of
	splitting alternative experiments.

    o   Added a mainExpName() getter and setter to remember the name of
	the main experiment.

    o   splitAltExps() will now store the chosen ref as the
	mainExpName.

    o   swapAltExp() now discards the promoted experiment from the list
	of alternative experiments in the output. It will also exchange
	the colData between the swapped experiments when
	withColData=TRUE. These changes assist in achieving
	reversibility of the output.

    o   Added applySCE() to conveniently apply a function to the main
	and alternative Experiments.

    o   Added withDimnames= to reducedDim<-() and reducedDims<-().  If
	TRUE, these methods now emit warnings on observing incompatible
	row names in value.

    o   Respect any metadata passed in with value in reducedDims<-()
	and altExps<-().

    o   Added the reduced.dim.matrix class to preserve attributes
	inside the reducedDims during subsetting/combining.

    o   Setting withColData=TRUE in altExp() and altExps() will now
	prepend colData(x) to the output colData.

    o   Added withDimnames= to altExp<-() and altExps<-().  If TRUE,
	these methods now emit warnings on observing incompatible
	column names in value. Also added withColData=, which will now
	reverse the prepending in the getter if the left-most columns
	are the same as colData(x). (If not the same, a warning is
	emitted.)

                       Changes in version 1.12.0                        

    o   Added the rowSubset() function as a standard location for a row
	subset.

    o   Added colPairs() and rowPairs() to store pairwise information
	(e.g., for graphs).

    o   Added method specifications for S4Vectors compatibility.

                       Changes in version 1.10.0                        

    o   Removed deprecated modes for getting and setting reducedDims.

    o   Added the colLabels function as a standard location for column
	labels.

    o   Added more options to control sizeFactors behavior when size
	factors are missing.

                        Changes in version 1.8.0                        

    o   Added altExp() and related methods to get and set alternative
	Experiments.

    o   Added the splitAltExps() utility to create many alternative
	Experiments at once.

    o   Added the swapAltExp() utility to swap between main and
	alternative Experiments.

    o   Deprecated isSpike(), spikeNames() and related arguments for
	handling spike-ins, in favor of representing spike-ins as
	alternative Experiments.

    o   Deprecated type= in sizeFactors() and sizeFactorNames(), which
	were previously only required to store size factors for
	spike-ins.

    o   Internal change to the representation of reducedDims() to
	streamline subsetting and combining.

                        Changes in version 1.4.0                        

    o   Allow ... arguments to be passed to rowData() and colData().

    o   Added weights() methods for getting/setting observational
	weights.

    o   Added reducedDimNames<- method to set the names of reduced
	dimension slots.

    o   Added withDimnames= argument to reducedDim() and reducedDims().

    o   Exported getters and setters for internal metadata fields.

    o   Added developer instructions for making use of internal
	metadata fields.

                        Changes in version 1.2.0                        

    o   Added the clearSpikes() function to remove all spike-in
	information.

    o   Added the clearSizeFactors() function to remove all size factor
	information.

    o   Added the sizeFactorNames() function to query the available
	(named) size factor sets.

    o   isSpike() with an unknown spike-in set in type= will no longer
	throw an error, and will quietly return NULL.

    o   isSpike<- with type=NULL is deprecated in favour of
	clearSpikes() for removing existing spike-in information. All
	spike-in sets must also be explicitly named during assignment.

    o   Added the LinearEmbeddingMatrix class for storing
	dimensionality reduction with loadings.

                        Changes in version 1.0.0                        

    o   
	New package SingleCellExperiment, for representation of
	single-cell genomics data.