Re: CODEX2 run time error

[Jiang, Yuchao]

Hi,

It seems that all of your exons and targets were removed during the QC step and primarily due to low coverage. Is this whole exome sequencing? How large is your target?

You can bypass this by lowering the threshold of coverage (default set at a mean of 20 across all samples) however this may affect downstream performance.

Hope that this is helpful and let us know if you have further questions.

Yuchao

Sent from my iPhone

> On Jun 19, 2019, at 6:42 AM, "a.bha...@uliege.be" <a.bha...@uliege.be> wrote:
>
> Dear Yucaho
>
> I am working on in area of cancer genomics. I would like to predict CNV from WES datasets where I have normal and tumor samples. In your paper, you mentioned about the "case-control design with a group of negative control samples" . SO ,I am analysing it accordingly under case-control strategy. While running code, at quality check, I am getting following error.
>
>
> qcObj <- qc(Y, sampname, ref, cov_thresh = c(20, Inf),
> + length_thresh = c(20, Inf), mapp_thresh = 0.8,
> + gc_thresh = c(10, 90))
> Excluded 620587 exons due to extreme coverage.
> Excluded 20558 exons due to extreme exonic length.
> Excluded 37049 exons due to extreme mappability.
> Excluded NA exons due to extreme GC content.
> After taking union, excluded NA out of 747477 exons in QC.
> Error: logical subscript contains NAs
> Execution halted
>
>
> Could you please help me to fix this issue?
>
>
> kind regards
> Archana

[a.bha...@uliege.be]

Hi Jiang,

Thank you for your quick response. Yes its Whole exome sequencing data. I have normal and target WES separately (25 samples each). I will try to reduce the coverage and redo run.

kind regards
Archana

----- Mail original -----
De: "Jiang, Yuchao" <yuc...@email.unc.edu>
À: "a bhardwaj" <a.bha...@uliege.be>
Cc: cod...@googlegroups.com
Envoyé: Mercredi 19 Juin 2019 13:54:16
Objet: Re: CODEX2 run time error

[Jiang, Yuchao]

You are welcome. You may also consider combining certain regions. If you only have on average 10counts/exon under the null, it may be hard to call CNVs at high confidence.

Yuchao