3.3.2

• Fixed --yAxisLabel in plotProfile (issue #889)
• Fixed a small X-axis tick offset issue. This caused the location of tick marks in profile plots to be shifted to the left by 0.5 to 1 bin. This was generally not notable, only really appearing when very few bins (e.g., 4) were used. The issue was mostly that the end tick would appear after the end of the plot, since its coordinate was the end of the bin. (issue #888)
• multiBamSummary and multiBigwigSummary no longer exclude small bins at the end of genomic chunks. multiBamSummary now has a --genomicChunkSize option in case users need to control the size of the genome used for multiprocessing for consistency. (issue #887)
• Added 4 new colormaps, which were copied from the seaborn project (issue #879). These are: rocket, mako, vlag, and icefire.
• Fixed an issue in the Galaxy wrapper of plotCorrelation where the X and Y.
• Fixed an issue with the --Offset option, where a single negative value wouldn't include only a single position, but rather that base through the end of the read. (stems from issue #902)
• Clustered output from plotHeatmap and plotProfile now allow computing the silhouette score of each row. This is printed in the returned BED file as the last column.

3.3.1

• Fixed --plotNumbers not working in plotCorrelation. This was issue #838.
• Fixed compatibility with matplotlib 3 and restrict to at least that version.
• The Y-axis labels should once again appear in both plotHeatmap and plotProfile (issue #844). This was related to the previous point.
• Testing is no longer performed with python 2.7, which will reach end of life in a couple months.
• Various documentation updates (issues #868, #867 and #851).
• Increased support for BED files with track header lines (issue #866).

3.3.0

• plotCoverage now has a --BED option, to restrict plots and output to apply to a specific set of regions given by a BED or GTF file or files (issue #829).
• plotCoverage now has a --DepthSummary option, which produces a summary similar to GATK's DepthOfCoverage (issue #828).
• plotCoverage is now able to compute coverage metrics for arbitrary coverage thresholds using multiples of the -ct option (e.g., -ct 0 -ct 10 -ct 20 -ct 30).

3.2.1

• Changed a bug in estimateReadFiltering where the estimated number of filtered reads was typically too low.
• Made an internal change that should drastically reduce the memory requirements of many tools. This slightly increases run time, but as the resulting resource usage is much more attractive this is judged worthwhile.
• An informative error message is now produced with bamCoverage if RPGC normalization is requested but no effective genome size is provided (issue #815).
• Fixes some issues with y-axis scaling (issue #822)

3.2.0

• Added access in the Galaxy wrapper to the --labels option in most tools (issue #738)
• Added the std plot type to plotProfile in Galaxy (issue #782)
• bamCompare now has a --skipZeroOverZero option to allow skipping bins where both input files lack coverage (issue #785)
• bamCompare and bigwigCompare can now take two pseudocounts, in case you want a different value for the numerator and the denominator (issue #784)
• multiBamSummary now has a --scaleFactors option, which computes scale factors in the same manner as DESeq2 to a file. Note that the produced scaling factors are meant to be used with bamCoverage. If you want to use them directly in DESeq2 (or a similar package) you will need to invert them (take 1/scale factor). (issue #800)
• Fixed an issue with large numbers of samples and small genome sizes sometimes causing nothing to be processed. (issue #801)

3.1.3

• Added the --legendLocation option in the Galaxy wrappers for plotProfile and plotHeatmap
• More thoroughly checked that output files can be written (issue #764).

3.1.2

• Added a --markers option to plotPCA, courtesy of @sklasfeld.
• computeMatrixOperations rbind now properly supports multiple region groups (issue #742)
• Fixed the usage of --xRange and --yRange with plotCorrelation (issue #709)

3.1.1

• Fixed the --outFileNameData option in plotProfile when computeMatrix reference-point --referencePoint center was used. This caused an error previously. (issue #727)
• RPGC normalization and the --scaleFactor option in bamCoverage are no longer mutually exclusive.
• Increased the default plot width in plotPCA (issue #738)

3.1.0

• The --centerReads option in bamCoverage is now compatible with --Offset (previously --centerReads was silently ignored if --Offset was specified). (issue #693)
• bamCoverage and bamCompare now have an --exactScaling option. Instead of using a random sample of alignment to compute the scaling factor, this causes all reads in the file to be used. This is significantly slower, but helpful in situations where reads that should be excluded clump together on the genome (i.e., when sampling based on location is likely to be inaccurate).
• plotCorrelation --whatToPlot scatterplot now has --xRange and --yRange options rather than just --maxRange. (issue #709)
• computeMatrixOperations can now be used to change sample and group names.
• computeMatrixOperations can now filter rows by minimum and/or maximum value.
• --maxThreshold and --minThreshold are now more consistently honoured. (#702)
• Fixed region handling when using files on deepBlue (#700)
• Using --normalizeUsing RPGC with bamCompare will now result in a fatal error, rather than a simple warning and the settings being changed under the hood. (#718)
• Related to the last point, setting --normalizeUsing to anything other than None will result in an error unless --scaleFactorsMethod None is also used. This is to prevent people from accidentally getting unintended normalization.
• bamPEFragmentSize no longer exploids its memory use with multiple large BAM/CRAM files (#720). Many other tools will also benefit from this change.

3.0.2

• Fixed an issue regarding under sampling alignments in some cases with computing scaling factors. This was issue #690. The resolution isn't perfect, it's hard to know how many reads really need to be sampled for things like RNA-seq.
• computeMatrix now has a --verbose option. Setting this will drastically increase the verbosity of the messages sent to the screen. Only do this for debugging. --quiet will disable this completely (as well as all other messages printed to screen).
• Fixed handling of --sortUsing region_length in plotHeatmap. This now works properly for --referencePoint center and --referencePoint TES, where in the latter case the dashed line is drawn at the region start. The documentation has been updated to mention this. (issue #671)
• The reference point label specified by computeMatrix reference-point is now respected by plotHeatmap and plotProfile. So if you used computeMatrix reference-point --referencePointLabel center then 'center' will now appear as the tick label in your heatmaps and profiles automatically. (issues #606 and #683)
• Enabled using regions with a . in the chromosome name in the Galaxy wrappers (issue #692)