Changes in version 1.21.1

   o replaced seq_name method by seqnames
     (due to deprecation of seq_name generic)

Changes in version 1.3.11

   o When reducing overlapping reads, the read counts are now directly
     taken into account as weights for deriving the consensus sequence.
     To that aim, a custom function "weightedConsensusMatrix" is now
     used in place of "consensusMatrix" from Biostrings

Changes in version 1.3.8

   o Several minor changes including corrections to the documentation.
     For plotting the aligned reads, the user can now specify custom
     colours for certain types of annotated feature. See
     ?plotAligned
     for an example and details.

Changes in version 1.3.6

   o New function "countReadsAnnotated" to sum up aligned reads per
     category of genome feature (i.e. gene, ncRNA, etc.).

   o New function "medianByPosition" that allows for a
     memory-efficient computation of the median base quality
     at each nucleotide position in the reads

Changes in version 1.3.5

   o Changed implementation of 'sample' method. The 'Rle' objects
     used before were not suitable for large AlignedGenomeIntervals
     objects. 'sample' now uses the base::sample.int function. As
     a consequence only sampling with replacement is possible for
     the moment.

   o Added a logical argument 'both' to the 'fracOverlap' function. 
     IF TRUE (default), both overlapping partners must be involved
     to at least a proportion of 'min.frac' in order to report the
     overlap; if FALSE, it suffices that one of the partners meets
     the 'min.frac' requirement.
 
Changes in version 1.3.4

   o Added an argument 'method' for the 'reduce' function. This 
     argument supersedes the boolean argument 'exact', use
     'method="exact"' in place of 'exact=TRUE'. Additional options
     are 'method="same5"' and 'method="same3"' to combine intervals
     that have the same 5' or 3'end, respectively.

   o Added a "fasta" format option to the export method for
     AlignedGenomeIntervals objects.

Changes in version 1.3.2

   o Added slot 'chrlengths' to S4 class 'AlignedGenomeIntervals'
     This optional slot stores a named integer vector of
     chromosome lengths. If present, it is used by 'coverage' and
     other methods, instead of obtaining those chromosome lengths
     from the annotation package indicated by the 'organism' slot.

   o Added an "export" method for objects of classes 
     "Genome_intervals" and "Genome_intervals_stranded"

Changes in version 1.1.15

   o Certain functions (reduce, interval_overlap, fracOverlap)
     got an extra argument 'mem.friendly' that if set to TRUE
     (default: FALSE) leads to the data broken down into
     chromosome-wise batches for the computations, and thus
     for a "memory-friendlier", but slower, version of these
     computations.

Changes in version 1.1.12

   o Added a new function 'agiFromBam' which uses Rsamtools
     for creating AlignedGenomeIntervals objects directly
     from BAM files

Changes in version 1.1.6

   o Added an 'id' slot to AlignedGenomeIntervals objects, which
     allows users to assign a custom identifier to each interval

Changes in version 1.1.5

   o Added a 'sample' method for AlignedGenomeIntervals objects
     Consists of drawing a random sample of n of the aligned reads
     and returning the AlignedGenomeIntervals object defined by
     these aligned reads

Changes in version 1.1.3

   o Added a NEWS file

   o added a 'sort' method for AlignedGenomeIntervals to sort them 
     by chromosome number > start position > end position

   o added a 'hist' method for AlignedGenomeIntervals which creates
     a histogram of the lengths of the reads aligned to the intervals
     Like the 'hist' function from package graphics, it returns an
     object of class 'histogram' and accepts the argument 'plot=FALSE'

Changes in version 1.1.2

   o in function 'fracoverlap', the argument 'min.frac' now applies
     to both intervals and specifies the minimum required fraction of
     each of the two interval widths by which two intervals should 
     overlap in order to be marked as truly overlapping (before: only
     applied to the shorter of the two intervals)