Re: [svtoolkit-help] Genome STRiP - mismatched read pair insert size error

[Bob H. <han...@br...>]

Hi,

Yes, this problem comes up occasionally - I don't know whether it is 
indel realignment or some other tool (but a corner case in indel 
realignment is a good guess).
I find that I see it a few times in any largish data set that comes from 
the bwa/GATK pipeline.

You can work around this problem by changing a configuration parameter 
select.validateReadPairs:false.
You can do this by changing the config file, or in newer versions of 
Genome STRiP the preferred way is to leave the standard config file 
unmodified and put
-P select.validateReadPairs:false on the command line.

This flag controls several consistency checks.
I can't recommend turning this off by default, since if you have bad 
input data you really do want to find out about it.

Best,
-Bob

On 10/9/13 12:30 PM, Anne-Katrin Emde wrote:
> Hello!
>
> I am running GenomeSTRiP on 50 genomes (bwa-aligned, GATK indel 
> realigned and quality recalibrated) and I am getting the following error:
>
> ##### ERROR MESSAGE: Mismatched read pair insert sizes for sample 
> 6837: [ {HS2000-910_287:1:1309:11124:98404    97      17  22252630     
>    0       100M    =       22251548  -983 
>  CTTTGAAGATTTCGTTGGAAACGGGATAATCTTCACAGAAAAGCTAAACAGAAGCATTCTCAGAAACTTCTTTGTGATGTTTGCTTTCAACTCACAGAGT 
> >@??>>?=>=??=6=?==???<6==>>>?>=??=>;><=>??==?>??<>=>9==>=?=><=>>>?;>><>@?>=?>=<>>?><=?><=5;?=><>>>>= 
>    X0:i:5  X1:i:0 
>  XA:Z:17,+22255009,100M,0;17,+22259766,100M,0;17,+22247875,100M,0;17,+22245496,100M,0; 
> BD:Z:NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 
>    MD:Z:100        RG:Z:6837       XG:i:0 
>  BI:Z:NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 
>       AM:i:0  NM:i:0  SM:i:0  XM:i:0  XO:i:0  MQ:i:47    XT:A:R}, 
> {HS2000-910_287:1:1309:11124:98404     145   17      22251548       
>  47      2M1D98M = 22252630        981 
> TTTGAGAGAGAAGCTTTGAAACACTCTTTTTCTAGAATCTGCAAGTGGACATTGGGAGGGCTGTGAGGTTTGTGGTGGAAAAGGAAATATCTCCACATAA 
>  @@>=?=?=?=>?<>=>?>?==>;=><>>>>><=><?=><=<=?>;======>>==<>====>;>=?=;??>;>=;===????==??=>=>=>>=<===@7 
> X0:i:1  X1:i:0  OC:Z:100M 
> BD:Z:NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 
>       RG:Z:6837       XG:i:0 
>  BI:Z:NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 
>    AM:i:0  NM:i:4  SM:i:37 XM:i:4  XO:i:0  OP:i:22251549   XT:A:U} ]
> ##### ERROR 
> ------------------------------------------------------------------------------------------
>
> So the insert size of the left and right mate don't agree. We think 
> that indel realignment might cause the mismatch in insert sizes. Has 
> this error message been reported before on bwa-aligned reads (+ GATK 
> realignment, recalibration)? Is there a way to prevent GenomeSTRiP 
> from crashing at those instances, e.g. printing a warning instead of 
> an error?  It only happens very rarely - on 50 high-coverage whole 
> genomes the error was reported 4 times.
>
> Thanks,
>
> Anne-Katrin
>
>
> --------------------------------------
>  Anne-Katrin Emde, Ph.D.
>  Bioinformatics Scientist
>  New York Genome Center
> ak...@ny... <mailto:ak...@ny...>
>  (917)-951-0167
> --------------------------------------
>
>
>
> ------------------------------------------------------------------------------
> October Webinars: Code for Performance
> Free Intel webinars can help you accelerate application performance.
> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from
> the latest Intel processors and coprocessors. See abstracts and register >
> http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
>
>
> _______________________________________________
> svtoolkit-help mailing list
> svt...@li...
> https://lists.sourceforge.net/lists/listinfo/svtoolkit-help