HiCExplorer icon indicating copy to clipboard operation
HiCExplorer copied to clipboard

hicCorrect with cool format

Open lldelisle opened this issue 5 years ago • 1 comments

Hi, when you use the cool format, the correct values are stored in the raw and when you apply the weights you get back to before correction... I think this is due to hicmatrix 13. Here is a minimal test to explain: I used the small_test3.cool that you can find (here)[https://github.com/lldelisle/pyGenomeTracks/raw/logInMatrices/pygenometracks/tests/test_data/small_test3.cool].

conda create -n hicexplorer3.5 hicexplorer=3.5
conda activate hicexplorer3.5

hicCorrectMatrix diagnostic_plot -m small_test3.cool --plotName diag3
hicCorrectMatrix correct  -m small_test3.cool --correctionMethod ICE --filterThreshold -2 5.0 -o ice3.cool
hicConvertFormat -m ice3.cool -o ice3_raw.cool --inputFormat cool --load_raw_values --outputFormat cool
# This generate a cool format with h5 extension... I will write another issue for that...
# hicCorrectMatrix correct  -m small_test3.cool --correctionMethod ICE --filterThreshold -2 5.0 -o ice3.h5
hicConvertFormat -m small_test3.cool -o small_test3.h5 --inputFormat cool --outputFormat h5
hicCorrectMatrix correct  -m small_test3.h5 --correctionMethod ICE --filterThreshold -2 5.0 -o ice3.h5
echo "[ori]
file = small_test3.cool
title = input
depth = 2000000
show_masked_bins = true
min_value = 0
max_value = 50

[ice]
file = ice3.cool
title = ice
depth = 2000000
show_masked_bins = true
min_value = 0
max_value = 50

[ice]
file = ice3_raw.cool
title = ice raw values
depth = 2000000
show_masked_bins = true
min_value = 0
max_value = 50

[ori]
file = small_test3.h5
title = input as h5
depth = 2000000
show_masked_bins = true
min_value = 0
max_value = 50

[ice_h5]
file = ice3.h5
title = ice when using h5
depth = 2000000
show_masked_bins = true
min_value = 0
max_value = 50


[x-axis]
" > test3.ini
pgt --tracks test3.ini --region chr1:0-200000 -o test6.png

test6

For info, I tested it with hicexplorer 3.4.3 and 3.4.1 and I always got the same output.

lldelisle avatar Jul 16 '20 08:07 lldelisle

I think the issue is this line which should not be here: https://github.com/deeptools/HiCExplorer/blob/c2ad8630c48e04f0ce4229d75d4b1766df79317c/hicexplorer/hicCorrectMatrix.py#L740

lldelisle avatar May 16 '25 10:05 lldelisle