Skip to contents

Changelog

Source: NEWS.md

wigglescout 0.19.0

Date: 04/10/2024

Features

  • New three functions plot_gr to plot columns on a precalculated GRanges: plot_gr_scatter, plot_gr_violin, plot_gr_density.
  • Removed uninformative default titles in plots.

wigglescout 0.18.1

Other

Restructuring of vignettes in smaller pieces of information, and deployment of documentation website with pkgdown.

wigglescout 0.18.0

Features

  • keep_canonical function makes use of GenomeInfoDb functionality and exposed as utility.

wigglescout 0.17.0

Features

  • Extra canonical parameter on bw_bins and plot_bw_bins_density, plot_bw_bins_scatter which allows to exclude all extra contigs on a reference genome.

  • bw_bins related functions now check if bigWig file references match to each other and to the tile genome provided, giving a, initial warning about potential references missing, but suppressing all the individual “Unable to summarize range” warning messages from rtracklayer summary calculation.

wigglescout 0.16.1

Bug fixes

  • Fixed issue importing formats different to BED files. Now the format guessing is left to rtracklayer and narrowPeak files are accepted. Fixes issue #98.

wigglescout 0.16.0

Features

  • plot_bw_bins_density function introduced to separate this type of figure from plot_bw_bins_scatter functionality. Use of density = TRUE in plot_bw_bins_scatter is soft-deprecated and will throw a warning.
  • Density plots tile layer is rastered to a resolution of 300dpi.

Other

  • Vignettes now include an example of a density plot.

wigglescout 0.15.6

  • Now scatterplots and jitterplots on violins are raster layers, with enough resolution for prints (300dpi). This considerably reduces the size of the figures and makes them more tractable on editing.
  • ggrastr is now a dependency.

wigglescout 0.15.5

  • Minor refactoring changes.
  • Fixed issue with tidyselect::where not being exported.

wigglescout 0.15.4

  • Fixed warnings from deprecations in ggplot - removed all appearances of aes_string and replaced those with the recommended .data$field notation instead. Additionally, incorporated legend.position = inside to the profile plots.
  • wigglescout now requires ggplot2 >= 3.5.0 due to these changes.
  • Look of figures improved by better positioning of legend, more adequate base fontsize and truncation of too long labels (>35 characters). Legend title in summary heatmaps is also shortened (background renamed as bg).
  • Verbose caption on heatmaps shows if the order of rows is the default (decreasing by row means) or user_provided.
  • Bug fix #68 - colors are now correctly assigned to labels in plot_bw_profile.

wigglescout 0.15.3

  • Testing refactor to exclude as much global variables as possible. Now some of them have been moved to fixture rds files that are loaded on demand, and also some of the necessary files generated on the fly per test function call and removed after with withr. This might generate a bit of overhead but improves legibility and maintainability of the tests.

wigglescout 0.15.2

  • Fixed #96: Inconsistency when bw_loci was summarized, normalized to background and remove_top > 0. Now loci selection for exclusion is done at the same time to both background and main samples.
  • Fixed deprecation warnings for tidyselect 1.1.0, 1.2.0, ggplot 3.4.0: aes_string replaced by aes as tidyselect now goes for strings instead of .data[[“string”]], as explained on: https://github.com/tidyverse/tidyverse.org/pull/600 This impacts ggplot2, and there is no easy fix, since .data[[col]] is also deprecated. All calls for dplyr::select that use vectors have been replaced by the relevant tidyselect::all_of, tidyselect::any_of use.
  • Extra deprecation warnings for ggplot 3.4.0: linewidth replaces size for line and rectangle functions.
  • The versions of the packages where this is deprecated are added on the DESCRIPTION file, since these are breaking changes for older versions.
  • Added a warning when heatmap colorscale is min == max. This can happen due to extremely sparse heatmaps. For consistency, if this happens it is plotted using raw colorscale (instead quantiles (0.01, 0.99), use (min, max)).

wigglescout 0.15.1

  • Reimplemented granges_cbind function to make use of dplyr::left_join instead, improving robustness of the approach. Ranges are also deduplicated before merging. In the case where a granges locus has more than one name, the final granges object will have as many rows as names.
  • .loci_to_granges now does not sort the sequences, since it is not necessary for having a proper merging and keeps consistency.
  • .loci_to_granges now gets rid of any column that is not a name, so it does not interfere with scoring naming.

wigglescout 0.15.0

  • Added bw_global_mean function. Takes a bigwig file and returns global mean coverage by taking the per-contig summary from rtracklayer, multiplying by contig length and dividing the sum by global length.
  • Added a scaling parameter that divides results by global mean calculated by bw_global_mean function, to simulate 1x genome coverage tracks. This affects all the calculating bw_ functions and the plotting functions: plot_bw_ but it is backwards compatible (default is not to do anything).
  • are assigned correctly. This affected only instances where BED file provided was not sorted. Might have been introduced recently, since older calculations are still correct.

wigglescout 0.14.0

  • Added a default_na parameter to all the functions. This makes use of rtracklayer summary function defaultValue parameter. Bins or loci that have no values are replaced by the given value instead. This is particularly useful with very sparse bigWig files, where .aggregate_scores might kill majority of rows due to NA values.

wigglescout 0.13.9

  • .aggregate_scores is robust to presence of NAs. Only rows with no NA values are aggregated. This affect results from plot_bw_summary_heatmap and bw_loci when aggregate_by is not NULL.

wigglescout 0.13.8

  • .bw_ranges refactored to make use of .fetch_bigwig function.

wigglescout 0.13.7

  • bw_profile bug fix: remove_top now ignores NA values for calculating rowMeans and quantile.
  • Documentation fix: Removed unnecessary packages from README.

wigglescout 0.13.6

  • Refactor plot functions to separate outlier filtering from plotting and reporting.
  • reshape2 dependency melt is removed in favor of tidyr pivot_longer.
  • Explicit imports of ggplot2 functions.
  • Better handling of decimal positions in verbose plots.
  • Remove soft-deprecated aes_string calls and replace with .data.
  • Row aggregation is now more precise in plot_bw_heatmap.
  • Missing tick 0 on heatmap on row aggregation fixed back. Replaced by 1 value.

wigglescout 0.13.5

  • Code style fixes:
    • All lines under 80 characters.
    • Replace sapply() -> vapply().
    • Replace calls to class() to compare with is(obj, “class”).
    • Replace T/F values with TRUE/FALSE.
    • Calls like warning(), stop() do not contain paste() calls or error keywords.
    • Indent 4 spaces all.
    • Plotting functions now dynamically match inner parameters. Tests for inner parameter calls are hence removed.

wigglescout 0.13.4

  • Added a NEWS.md file to track changes to the package.
  • Fixed DESCRIPTION to comply with Bioconductor’s BiocCheck:
    • Added biocViews terms.
    • Fixed inconsistencies in imports/suggests packages.
    • Expanded description of the package.
    • Correctly formated title of the tool.
  • Made all non-exported functions start with dots.
  • Added sessionInfo() to the end of Vignettes.
  • Added examples to all exported functions.