CHANGES IN VERSION 1.27.0
--------------------------

  o GenoSet objects now inherit from RangedSummarizedExperiment (RSE) rather than eSet. Use
    the RSE API rather than that for eSet (e.g. colnames rather than sampleNames and assays
    rather than assayData). All Biobase-related content has been removed including
    annotatedDataFrameFrom. This means that BigMatrix objects can no longer be added to
    eSet/ExpressionSet/GenoSet objects.


CHANGES IN VERSION 1.21.10
--------------------------

NEW FEATURES

    o calcGC gets a 'bases' argument and handles the presence or absence of 'chr'
    prefixes in a safer manner.

CHANGES IN VERSION 1.19.27
--------------------------

NEW FEATURES

    o rangeMeans for numeric gets a na.rm argument and uses 50% less time for 
    na.rm = FALSE ands 25% less time for na.rm=TRUE.

CHANGES IN VERSION 1.19.7
-------------------------

NEW FEATURES

    o The version introduces the rangeMeans family of functions, which are 
    closely related to IRanges::viewMeans family of functions. These 
    new functions take an RleDataFrame and an IRanges (or matrix) of row 
    ranges and compute summary stats. These are useful for summaries like 
    the average DNA copy number or coverage in a gene, by sample. These 
    functions differ from view* in that they are more time efficient and 
    they offer a 'simplify' argument to return a matrix, rather than a list.

CHANGES IN VERSION 1.19.0
-------------------------

NEW FEATURES

    o This version introduces the RleDataFrame class. This class extends 
    SimpleRleList and DataFrame. It stores a collection of like-length 
    Rle objects.  With the AtomicList features it behaves much like a 
    matrix. For example, log2(x) - 1 works. We also have row and column sums 
    and means. rangeSampleMeans functions like viewMeans.
    This object can be used to store 
    runs of data along the genome for multiple samples, like coverage or DNA 
    copy number. It can be used as an assayDataElement in a genoset (or any 
    eSet).

    o segTable and segPairTable are dramatically faster. Stacking data.frames 
    with segTable(RleDataFrame,stack=TRUE) argument is now instantaneous.

DEPRECATED AND DEFUNCT

    o Previously we used a DataFrame for storing collections of Rle objects. 
    This practice is now deprecated. Similarly, colMeans on a DataFrame is 
    now deprecated.

CHANGES IN VERSION 1.16.0
-------------------------

NEW FEATURES

    o cn2lr now has methods for vector, matrix, and DataFrame (of Rle) and 
    allows you to center your log2ratio values on 2 copies, or on a specified 
    'ploidy' for your sample.

    o The BAFSet and CNSet classes have moved from deprecated 
    to defunct. These classes only added getter/setter methods 
    for baf/lrr/cn. Since these only cover some possible 
    assayDataElements, it is better to use x[i,j,k], where k is 
    the name of an assayDataElement.
    
    o All RangedData-related things have now progressed to defunct. Please use 
    GRanges for locData and everywhere else. Since the genoset package provides 
    a common API for GenoSet, GRanges, and RangedData, I hope this will be a 
    simple change for everyone.
    
    o sampleNames and featureNames are now *un-deprecated*.  Feel free to use 
    them. They just call colnames and rownames, respectively. I have defined 
    my own rownames and colnames functions. The eSet ones seem to do a lot of 
    extra work, and the getter versions read from the first assayDataElement, 
    rather than pData and fData. I changed to the latter, so BigMatrix 
    assayDataElements will remain untouched until you really mean to access 
    them.
