None valies in Y axis of Hi-C data

[apaytuvi]

I upgraded HiCExplorer version and now hicPlotTADs outputs (with transform = log1p) the following axis:

[fidelram]

I can not replicate the problem with my matrices. Could you share privately the data with me?

[fidelram]

It is working for me:

the tracks.ini file contains:

[x-axis]

[hic Kc]
file = hicmatrix.h5
title = test
colormap = RdYlBu_r
depth = 20000000
transform = log1p

hicPlotTADs --tracks tracks.ini --region chr10:1000000-30000000 --outFileName test.png

Do you get any error message?

[apaytuvi]

I get:

{'chr6': 10, 'chr5': 11, 'chr4': 22, 'chr3': 17, 'chr2': 4, 'chr8': 8, 'chr13': 25, 'chr11': 3, 'chr10': 4, 'chrX': 8, 'chr18': 4}
found existing 116 nan bins that will be included for masking 
WARNING: bin size is not homogeneous. Median 1000000
time initializing track(s):
5.11754608154
Figure size in cm is 40 x 16.8372413793. Dpi is set to 72
setting min, max values for track 2. [hic Kc] to: 231.441678989, 11394.4467428

and this figure:

[fidelram]

What do you get if you type:

hicPlotTADs --version pip freeze | grep matplotlib

I assume that the track you are using is the same a me.

On Tue, Apr 4, 2017 at 4:42 PM, apaytuvi notifications@github.com wrote:

I get:

{'chr6': 10, 'chr5': 11, 'chr4': 22, 'chr3': 17, 'chr2': 4, 'chr8': 8, 'chr13': 25, 'chr11': 3, 'chr10': 4, 'chrX': 8, 'chr18': 4} found existing 116 nan bins that will be included for masking WARNING: bin size is not homogeneous. Median 1000000 time initializing track(s): 5.11754608154 Figure size in cm is 40 x 16.8372413793. Dpi is set to 72 setting min, max values for track 2. [hic Kc] to: 231.441678989, 11394.4467428

and this figure:

[image: test] https://cloud.githubusercontent.com/assets/11228225/24662570/b2288228-1955-11e7-9925-5faca6e02c76.png

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/maxplanck-ie/HiCExplorer/issues/83#issuecomment-291521862, or mute the thread https://github.com/notifications/unsubscribe-auth/AEu_1SQezd0nW8bbpV0-FC9rrF_6GVywks5rslbdgaJpZM4Myzlv .

--

Fidel Ramirez

[apaytuvi]

hicPlotTADs 1.7.2
matplotlib==2.0.0

[fidelram]

I will make branch such that you can get some debugging information.

[vivekbhr]

@apaytuvi it seems that this issue arises due to some weird combination of numpy-->matplotlib updates.. For now I have tried a fix in the develop branch. Can you try this?

first remove your current numpy and install numpy 1.12.0 :

pip uninstall numpy
pip install numpy # should install v1.12.0

Then (hoping no other conflicting version is installed in conda etc.), install hicExplorer from the develop branch

git clone https://github.com/maxplanck-ie/HiCExplorer.git
cd HiCExplorer
git checkout develop
python setup.py install -f # shall also check for new numpy/matplotlib versions

Then try to run hicPlotTADs again to see if the labels work now..

If it works we shall implement this properly and then merge in master

[apaytuvi]

Thank you!!