CHANGES IN VERSION 2.3.0
------------------------

NEW FEATURES

    o Refactored the bootstrap approach to improve memory usage and reduce
      calculation time.
    o Added a plot_metagene function to produce a metagene plot from a
      data.frame to avoid always having to rely on the metagene object.
    o Deprecated the range and bin_size params.
    o Added new sections in the vignettes: "Managing large datasets" and
      "Comparing profiles with permutations".

BUG FIXES

    o Removed params that no longer works with ggplot2 > 2.2.0.
    o Changed the seqlevels checks to match changes in GenomicAlignments.
    o Added a check to stop early and output clear error message when position
      in a GRanges is greater than the size of a chromosome.


CHANGES FOR VERSION 2.2.0
------------------------
    o Added checks to avoid producing identical matrices or data frame when the
    parameters are still the same after first function call.
    o Splitted the analysis in multiple (optionnal) intermediate steps
    (add_design, produce_matrices and produce_data_frame).
    o narrowPeak and broadPeak format is now supported.
    o Added multiple getter to access metagene members that are all now
    private (get_params, get_design, get_regions, get_matrices,
    get_data_frame, get_plot, get_raw_coverages and get_normalized_coverages.
    o Added the NCIS algorithm for noise removal.
    o Replaced the old datasets with promoters_hg19, promoters_hg18,
    promoters_mm10 and promoters_mm9 that can be accessed with
    data(promoters_????).
    o Added flip_regions and unflip_regions to switch regions orientation based
    on the strand.