2.4.1

• Setting --zMin to the same value as --zMax, whether intentionally or because the --zMax value computed by deepTools happens to be now larger than the desired value, will result in the maximum value in the dataset being used (internally, --zMax gets set to None).
• Scale factor is now set to 1 in bamCoverage if no normalization is used. The fact that this wasn't being done previously was a bug.
• Fixed a bug (#451) affecting BED files with a deepTools_group column that caused a problem with --sortRegions keep in computeMatrix.
• Fixed a bug where some matrices produced with computeMatrixOperations cbind would result in the right-most samples sometimes getting squished due to having ticks outside of their graph bounds. Ticks are now scaled if they don't match the data range (issue #452).
• In plotFingerprint, the number of reads per-bin are no longer used. Instead, the sum of the per-base coverage (or signal if bigWig input is used) is used. This leads to more similar metrics produced by us and others regarding things like Jensen-Shannon metrics. For those just interested in the plots, there's little effective change here.