CODEX2 - paired sample coverage

23 views
Skip to first unread message

Wipfler, Matthew

unread,
Jul 31, 2019, 10:51:36 AM7/31/19
to cod...@googlegroups.com

Hello CODEX2 devs, 


I have been looking into CODEX2 for paired tumor-normal CNV calling and had a couple questions. We have been testing CODEX2 with five WES tumor/normal samples sequenced in the same batch at depths 300x/100x. After following the demo and making adjustments, we ran CODEX2 with the normal samples as the negative controls and used the standard segmentation for output. 99,804 calls were made for the five tumor samples; 62,515 for the five normal. 20,000 CNVs per tumor sample seems excessive. From looking at the called regions in IGV, a lot of tumor calls had close to the 3:1 ratio of sequencing depth with little variation in depth distribution. Additional filtering will be necessary of course.


1) Do the latent factors adjust the disparity between tumor/normal sequencing depths? It is common to have samples with a 300x/100x or even 500x/100x.


2) Would normalizing samples with lower coverage than tumor samples affect the calls?


3) The README states normalization would have a higher impact than segmentation. Would it be worth a shot to see the impact of using the paired tumor normal segmentation script?


Most CNV callers we have been testing do not explicitly say if coverage differences are adjusted for.


Any insight or clarification would be appreciated,

Matt Wipfler


Wipfler, Matthew

unread,
Aug 1, 2019, 5:01:26 PM8/1/19
to cod...@googlegroups.com

Hello again,


Apologies, I have the found the mistake in my modified code. Please ignore the prior email.


Matt W


From: Wipfler, Matthew
Sent: Wednesday, July 31, 2019 9:51:32 AM
To: cod...@googlegroups.com <cod...@googlegroups.com>
Subject: CODEX2 - paired sample coverage
 

Yuchao Jiang

unread,
Aug 7, 2019, 12:32:18 AM8/7/19
to Wipfler, Matthew, cod...@googlegroups.com
Glad that the problem was solved. Please feel free to let us know if you need further assistance.

Cheers,
Yuchao


-- 
You received this message because you are subscribed to the Google Groups "CODEX/CODEX2 for CNV detection" group.
To unsubscribe from this group and stop receiving emails from it, send an email to codex2+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/codex2/CY4PR04MB119129D58479C5FFD630E6D1EDDE0%40CY4PR04MB1191.namprd04.prod.outlook.com.

Reply all
Reply to author
Forward
0 new messages