scalebar for hic matrix gets cutoff

[rikrdo89]

When plotting a hic matrix (from a cool file) , using the default settings from the make_tracks_file command, the scalebar gets cutoff on the left. The default transform is log1p, and the values of the scale bar with this transform are expressed as multiples of 10s (n x 10), which makes the label go off the plotting area. Is there a way to assign more plotting area to the left? or to change the notation of the scale bar to avoid using x 10?

Interestingly, when I use log transform, the matrix scale bar that gets plotted does not have any numbers. (is there a way to add them?)

With log1p:

With log:

Using pyGenomeTracks 3.8

[lldelisle]

Hi, Indeed, this is annoying.

1. a. For the labels that goes out of the image. This is something that I know and I would like to fix and I drafted a PR here: https://github.com/deeptools/pyGenomeTracks/pull/292 The trick I use is to export as pdf and open with inkscape so I can get back the labels but I know this is annoying and I hope this will be fix in the 4.0. Unfortunately, for the moment, I do not find the time to change it...
2. b. For the scientific notation, are you sure it is the log1p? Would you mind to give me the line of the log which says the min-max value? So I can have a look.
3. I am also aware of this and this is an old piece of code that I would like to change in 4.0. When choosing log or log1p, it forces to use 1, 2, 5, 10, 20, 50 but the developers did not think that we could have values below 1. So when this is the case you get no tick... Sorry.

[shRDJr]

Hello, I encountered the same problem as you. After converting my cool file to h5 format using hicConvertFormat and then drawing, the problem still occurred. Could you please tell me what the reason is?

[lldelisle]

Hi @shRDJr , Is your problem the same as described above? If it is the truncation, one trick I found is to play with --plotWidth and --trackLabelFraction. For example you can put: --trackLabelFraction 0.6 --plotWidth 30. It will make a super wide image but at least the y axis should not be cut.