AssertionError in "diagram" with WGS-level gene density

[etal]

AssertionError in "diagram" with WGS-level gene density

[mjafin]

There's another issue I could use help with in plotting,:

Traceback (most recent call last):
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/bin/cnvkit.py", line 9, in <module>
    args.func(args)
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/cnvlib/commands.py", line 685, in _cmd_diagram
    cnarr = CNA.read(args.filename) if args.filename else None
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/cnvlib/diagram.py", line 70, in create_diagram
    drawing = build_chrom_diagram(features, chrom_sizes, probe_pset.sample_id)
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/cnvlib/diagram.py", line 116, in build_chrom_diagram
    return bc_organism_draw(chr_diagram, title)
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/cnvlib/diagram.py", line 163, in bc_organism_draw
    sub_component.draw(cur_drawing)
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/Bio/Graphics/BasicChromosome.py", line 281, in draw
    sub_component.draw(cur_drawing)
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/Bio/Graphics/BasicChromosome.py", line 426, in draw
    self._overdraw_subcomponents(cur_drawing)  # Anything on top
  File "/group/ngs/src/bcbio-nextgen/latest-devel/rhel6-x64/anaconda/lib/python2.7/site-packages/Bio/Graphics/BasicChromosome.py", line 680, in _overdraw_subcomponents
    assert 0 <= start <= end <= self.bp_length

[etal]

I merged a pull request that should fix this. Can you try it now?

[mjafin]

@etal thanks, I managed to circumvent the initial problem but the second one above still bugs me. Is it maybe because it's a whole genome analysis and I'm trying to include the cnr file in the diagram command and it gets too crowded? Without the cnr file the diagram gets produced fine (but without gene labels)

[etal]

Do the alternative contigs in the .cnr file (chr11_gl000202_random, chr17_ctg5_hap1, chr17_gl000203_random) also appear in the .cns file? I'm guessing they don't, any in any case those contigs probably don't have usable copy number information on them. Can you try deleting those rows (i.e. those contigs) from the .cnr file and retrying the plots?

[mjafin]

I removed all the alternative contigs prior to getting the second error. Thanks for looking into this, again. I'll do some more testing tomorrow if it's something to do with the fact that we have sometimes overlapping genes on one row as comma separated lists.

[etal]

Sorry, I misunderstood what you did. If this is whole-genome data, then it could be an overcrowding issue where a feature gets squeezed to a size of less than a pixel. What averge bin size did you use to split the "targets" for your WGS dataset?

You could also try the "--short-names" option when processing the targets in order to resolve some of the overlapping gene names.

If you can link to any of the .cnr files you've created I can take a closer look.

[mjafin]

Hi @etal, I was calling cnvkit from within bcbio and it uses the following settings:

--antitarget-avg-size 1000 --antitarget-min-size 75 --target-avg-size 1000

[etal]

I haven't seen this error in a while, and testing with a WGS-density snippet of chr17 works fine for me. Maybe something was fixed in CNVkit or Reportlab in the interim. But please reopen if you see it again.