                        Changes in version 3.8.0                        

  - Addes sorting to the heatmap so reads appear from most to least
    methylated. This is done by calculating the average methylation
    across the region for each read and sorting by this value.

                        Changes in version 3.4.0                        

  - Fixed modbam_to_tabix() to not delete existing directories when
    provided as the output file. Previously it would delete the
    directory and create a file with the same name, now it will fail if
    the output is a directory.
  - Fixed gene annotation labels not showing up when symbol column is a
    factor.
  - Updated get_exons_t2t() to use character columns instead of factors.

                        Changes in version 3.2.0                        

  - Added preliminary support for importing modkit data using
    create_tabix_file().

                        Changes in version 3.1.0                        

  - Fixed parsing error for BAM files leading to an extre site to be
    called whenever skip width is greater than 0.
  - Added functions get_cgi_*() to get CpG islands annotation for mm10,
    GRCm39, hg19, and hg38.
  - Changed colour palette for heatmaps to have a less bright yellow on
    the low end.
  - Changed default NanoMethViz.site_filter value to 3. This filters out
    any sites with less than 3 coverage.
  - Changed BAM file parsing behaviour when parse flag "." or "?" is not
    set to default to "?". This is against SAM spec but follows the
    behaviour of IGV and older PacBio datasets.

                        Changes in version 3.0.0                        

  - Breaking change to smoothing strategy in plot_gene(), plot_region(),
    and plot_granges() to use weighted moving mean instead of loess.
    This deprecates the span argument in favour of smoothing_window
    which is defaulted to 2000 bases.
      - Smoothing for various plotting functions was previously
        performed using loess smoothing, this performed locally weighted
        linear estimation to create a smoothed line, the span argument
        controlled the proportion of data used in this smoothing. This
        parameter was difficult to tune because under a fixed span, the
        smoothed line became flatter as the plot region grew larger.
        Internally, NanoMethViz dynamically calculated a span that
        changed inversely proportional to the width of the plotting
        region, decreasing the span as the plot region grew. However the
        calculated span was invisible to users, and it unintuitive to
        users how to set a span to change the appearance of the plot.
      - Changing the smoothing method to a weighted rolling mean lead to
        the new smoothing_window argument which represents the window
        size in bases from which data is used for smoothing around each
        point. This serves the same purpose as the dynamic calculation
        done previously, but is set more explicitly and should be more
        intuitive for users. The default is always 2000, and can be
        increased to increase smoothness and decreased to decrease
        smoothness.
  - Breaking change to the appearance of plot_gene() plots, previously
    the isoform annotation would be restricted to only the gene of
    interest. It is now changed to follow the same behaviour as
    plot_region() whereby all isoforms in the region are plotted.
  - Breaking change to the default plotting options for plot_region(),
    plot_gene() and plot_grange() to plot heatmap by default.
  - Possible breaking change to query_methy(simplify = FALSE), it will
    now return a list that is the same length as the number of regions
    queried, where it previously returned nothing if a particular
    sequence was missing from the tabix.
  - Added gene_anno argument to plot_region() and plot_granges() to
    control whether gene annotation is plotted.
  - Added plot_violin() function for creating violin plots for samples
    over specific regions.
  - Added check to remove hard-clipped reads because they may not have
    matching mod strings.
  - Changed gene annotation to always put label on visible isoforms,
    previously labels are plotted at the center of isoform.
  - Changed the order of columns when querying from ModBam to be the
    same as when querying from Tabix, with readname at the end instead
    of being the second column.
  - Changed query_methy() to always return same length output as the
    query when simplify = FALSE.
  - Fixed memory leak in bam parsing due to out of bounds access.
  - Fixed crash when CIGAR doesn't match length of SEQ.

                        Changes in version 2.6.0                        

  - Added preliminary modBAM file support.
  - Changed rug plot to appear under other geoms. This helps with
    visibility of data when methylation values are close to 0.
  - Changed heatmap alpha from 0.5 to 1, line width from 1.0 to 1.2 and
    line colour from black to darkgrey.
  - Changed x-axis limits on plots to be controlled using
    coord_cartesian instead of scale_x_continuous. Plots should now
    accurately represent data around the boundaries.

                        Changes in version 2.4.0                        

  - Fixed plot_region_heatmap() producing the wrong plot when a factor
    is used for the chromosome.
  - Fixed nanopolish and f5c import positions being off by 1.
  - Fixed broken samples() setter for NanoMethResult.
  - Added plot_agg_genes() function as a shorthand for
    plot_agg_regions(x, exons_to_genes(exons(x))).
  - Added the ability to interrupt methy_to_bsseq() calls.
  - Added handling for NanoMethResult objects in filter_methy(). If
    NanoMethResult is used as input, then NanoMethResult is invisibly
    returned as output.
  - Added black outlines to exons in annotation to distinguish
    contiguous segments for features like tandem repeats.
  - Added line_size argument to plot_gene(), plot_region() and
    plot_granges() plots for adjusting line size.
  - Added subsample argument to heatmap plots, default 50. This reduces
    the number of rows shown the plot to the specified amount.
  - Added get_exons_mm10(), get_exons_hg19(), and get_exons_hg38() as
    replacements for get_exons_mus_musculus() and
    get_exons_homo_sapiens().
  - Changed heatmaps to no longer plot samples that are absent from
    sample annotations.
  - Changed heatmap labels to appear on the right rather than on top.
  - Changed heatmap alpha from 0.33 to 0.5.
  - Changed arrows in exon connectors to appear in the middle as open
    arrow instead of at the end as closed arrow.
  - Changed default X axis labels to be rescaled to appropriate
    SI-style. e.g. Kb, Mb, Gb.

                        Changes in version 2.2.0                        

  - Added heatmap argument to plot_gene(), plot_region() and
    plot_granges(). This adds a read-heatmap to the plot.
  - Added cluster_regions() function to perform k-means clustering on a
    table of genomic regions based on methylation profile.
  - Added median averaging method for trends in plot_gene(),
    plot_region() and plot_granges(). This can be changed using the new
    avg_method argument, default is mean.
  - Added filter_methy() function to create a filtered methylation file.
  - Added region_methy_stats() to obtain average methylation fractions
    of specific regions.
  - Added methy_to_edger() direct conversion wrapper around
    methy_to_bsseq() and bsseq_to_edger().
  - Added palette argument to plot_gene(), plot_region() and
    plot_granges() to allow custom colour palettes.
  - Fixed bsseq_to_edger() failing when regions argument was used.
  - Fixed heatmaps not staying in a single column when more than 2
    groups were present.

                        Changes in version 2.0.0                        

  - Major changes to plot_agg_regions().
      - Features of plot_agg_regions() and
        plot_agg_regions_sample_grouped() merged into one interface.
      - Regions now specified using single table.
  - Changed plot_regions() default window proportion to 0.
  - Changed default theme from theme_bw() to theme_tufte().
  - Added Megalodon data import instructions to "Importing Data"
    vignette.
  - Added scico palette defaults for heatmaps. These are colourblind
    friendly.
  - Added check for 0 length queries which would cause program to hang
    indefinitely.
  - Added setters for NanoMethResult attributes methy, samples and
    exons.
  - Added MDS and PCA plots.
  - Added vignette for using external annotation and dimensionality
    reduction.
  - Added binary thresholding for plot_gene(), plot_region() and
    plot_agg_regions().
  - Added regions argument to bsseq_to_edger() to calculate aggregate
    counts over features rather than per site.

                        Changes in version 1.1.4                        

  - Added palette argument to aggregate plots
  - Added exons_to_genes() function to convert exon annotation to gene
    annotation
  - Added plot_granges_heatmap() function to use GRanges for plotting
    heatmaps

                        Changes in version 1.1.3                        

  - Fixed group handling for list region input in plot_agg_regions()
  - Fixed unused window size argument in plot_region_heatmap()
  - Fixed error when reads overlap in name and position for internal
    function StatLM()

                        Changes in version 1.1.2                        

  - Changed example dataset exon annotations from all genomic exons to
    just those contained in data.
  - Fixed methylation heatmap to no longer be hard coded for Peg3.
  - Added plot_region_heatmap() as analogue to plot_region().
  - Fixed plot_agg_regions_sample_grouped() to use group column of
    NanoMethViz::samples(x) rather than haplotype.
  - Added unit tests.

                        Changes in version 1.1.1                        

  - Added methylation heatmap via plot_gene_heatmap().
  - Fixed gene_anno() in plot_gene() for argument so FALSE actually
    turns off gene annotation.
  - Added warning for cpp11 versions <0.2.5 which may cause memory
    crashes when trying to import methylation data.
  - Added cpp11 version dependency to address tidyverse/readr#1145.
  - Added query methylation by gene using query_methy_gene().

                        Changes in version 1.0.0                        

  - Initial Bioconductor release.