pindel2vcf error: unknown argument: –so about pindel

[liangkaiye]

Have you got a somatic parameter there?

Outlook for Androidhttps://aka.ms/ghei36

On Thu, Sep 8, 2016 at 9:43 PM +0800, "Maise B" <[email protected]mailto:[email protected]> wrote:

I am using Pindel to identify indels from tumour vs control. I would like to get the somatic p_values for the INDELs called using -so.
(-so/--somatic_p compute somatic p value when two samples are present, assume the order is normal and tumor. (default false))

my code:

pindel2vcf -P ../pindelfile -r ../ref.fasta -R hg19 -d 20160905 -so TRUE -v ../outputfile.vcf

I keep getting the error: unknown argument: –so

Can someone please help with this?
Many thanks

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHubhttps://github.com//issues/51, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AB9s-5-34EXSvZdUSGoUS3JRfLItt_XOks5qoBEHgaJpZM4J4AZX.

[MaiseB]

can you please clarify?

[MaiseB]

I am now no longer getting the aforementioned error. As mentioned above, I am trying to call indels/SVs from a tumour/normal pair (whole exome sequenced).

Can you please clarify how -so parameter is used? do I write -so true? or TRUE? or 1? The documentation is not clear how to change from the default setting of false.

The documentation also states that the order needs to be normal then tumour. Is that in the config file when running pindel to generate the SI/D/SV etc variants?

[EWLameijer]

Dear all,

as far as I can gather by analyzing the code, I do not think that the -so option works at the moment. I'll disable the option for now so that there won't be more confused/frustrated people, until such a time as the author of the -so code can either explain how it works and give example files so I and users can test it, or I (or Kai, or someone else) can get the code working.

Best regards,

Eric-Wubbo

[JohnMCMa]

From what I get from the current code, the --somatic_p parameter is a boolean parameter, and so should be used in the same way as -h. Looking at the code, this flag attempts to calculates a Fisher's p-value of ref frequency between the first and second samples (although the normal-tumor order doesn't matter, unlike what the doc says).

I'm checking now whether it works in the way I expect...

[JohnMCMa]

OK, I think I see the issue mentioned by @EWLameijer. When -somatic_p is triggered the program writes the aforementioned Fisher's p-value to the INFO string, but there was never a tag declared for this data. As a result a VCF line with that flag looks like the following; the p-value is the ;1 at the end of INFO:
GL000192.1 71722 . AT A . . END=71723;HOMLEN=8;HOMSEQ=TTTTTTTT;SVLEN=-1;SVTYPE=DEL;1 GT:AD 0/1:4 0/0:0
I'm now writing a fix for this issue right now.

[JohnMCMa]

I have written #69 to handle several issues involved in --somatic_p, which turns out to be more problem-laded than I expected.