plotHeatmap: Segmentation fault (core dumped)

[grahams...@gmail.com]

Hello,

I am unable to plot a portion of my data using the plotHeatmap command. I am only encountering a problem with a single region file. I have processed many others from the using the same script without issue. I can successfully generate the matrix but I then run into a segmentation fault.

The problem region file contains ~27,000 bed entries. I've broken this file into chunks and the commands work fine.

I am happy to post my matrix somewhere if it would help the troubleshooting.

Pertinent details:
python -c "import matplotlib as mpl; print mpl.__version__"
1.5.3

Commands
computeMatrix reference-point \
-R test.bed \
-S \
/data/reddylab/projects/GGR/data/chip_seq/quantified_read_counts/iter0/BCL3.t1.ctrl_subtracted.rpm.bw \
-out test.gz \
--missingDataAsZero \
--binSize 1 \
-b 10000 \
-a 10000 \
--numberOfProcessors max

plotHeatmap \
-m test.gz \
-out test.png \
--heatmapHeight 20 \
--heatmapWidth 7 \
--refPointLabel Summit \
--plotTitle "test1" \
--sortRegions descend \
--sortUsing sum \
--colorList 'white, black' \
--dpi 100 \
--verbose

[grahams...@gmail.com]

Forgot to post Python verion: Python 2.7.12

[Devon Ryan]

Yes, please post the matrix somewhere. It's incredibly unusual for
python to core dump. What version of deepTools do you have installed?
What happens if you update matplotlib to something more recent?

--
Devon Ryan, PhD
Bioinformatician / Data manager
Bioinformatics Core Facility
Max Planck Institute for Immunobiology and Epigenetics
Email: dpry...@gmail.com

[grahams...@gmail.com]

Here is the matrix: http://trackhub.genome.duke.edu/reddylab/gdj/test.gz

I am using deeptools 2.4.1. I can try updating matplotlib but I'm confused why only one bed file/region file fails..

Thanks
GDJ

[Devon Ryan]

I don't get a segfault, but I do get a memory error when I try this on
my desktop computer. It turns out that you have a very large matrix of
~550 million values. You just need a LOT of memory to actually plot
that given all of the sorting and other things that matplotlib is
trying to do. I'm attaching the results of the command you tried,
which took ~50GB of memory. In general, I would encourage you not to
bother with 1 base bins for things like this unless you plan on making
the figure the size of a highway billboard (you won't be able to see
the bins anyway).

Devon
--
Devon Ryan, Ph.D.
Email: dpr...@dpryan.com
Data Manager/Bioinformatician
Max Planck Institute of Immunobiology and Epigenetics
Stübeweg 51
79108 Freiburg
Germany

> --
> You received this message because you are subscribed to the Google Groups "deepTools" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to deeptools+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.