Go Back   SEQanswers > General

Similar Threads
Thread Thread Starter Forum Replies Last Post
how to check copy number from certain gene cliff Bioinformatics 2 07-14-2014 03:52 PM
exome copy number and LOH rcorbett Bioinformatics 20 11-26-2013 01:22 AM
CNANorm: A normalization method for Copy Number Aberration in cancer samples. stefanoberri Literature Watch 25 01-07-2013 02:01 AM
tools to detect copy number variants nans_bn Bioinformatics 2 01-12-2012 05:25 AM
RNA-Seq: Normalization strategies for microRNA profiling experiments: a 'normal' way Newsbot! Literature Watch 0 08-13-2010 09:00 AM

Thread Tools
Old 01-20-2012, 09:57 AM   #1
Location: Cambridge, MA

Join Date: Oct 2011
Posts: 11
Default Copy number normalization in capture experiments?


I'm trying to calculate rough copy number changes from our targeted (and generally deep) sequence of 3700 exons. However, I have a pretty substantial GC bias issue occurring in some of my samples. So far, I've just done rough bin normalization rather than a loess normalization. It seems that I can bring my data into the correct range, but that there's still a variance issue. I like what is suggested for corrections in the CNAnorm publication, but I feel unsure of how well it would work for such deep coverage localized to so few locations. Does anyone have any suggestions or thoughts?

anjulka is offline   Reply With Quote
Old 01-23-2012, 09:21 AM   #2
Location: British Columbia

Join Date: Jan 2012
Posts: 49

I have had numerous conversations with the author, he states that although it was designed for low coverage, he has been running it on 30X with no issues. I can back this statement up and say the same. It does work but its a bit trickier, try to set the window size appropriately
dmacmillan is offline   Reply With Quote
Old 01-23-2012, 01:07 PM   #3
Location: Cambridge area, UK

Join Date: Jan 2010
Posts: 35


I am pleased to see that CNAnorm is considered also outside the task it was designed for.
We are using CNAnorm for high coverage data and we do have a couple of capture experiments we'll be looking into.

Here some thoughts and observation that might help you.

- The most important aspect is to have the capture for a matched normal. It will partially balance the bias due to the capture.

- The capture is affected by CG content and some region with very high or very low GC content are not captured at all.

- Could be a good idea to "clean" your bam files so that reads are only from the expected target regions. You will get rid of all the windows with few off-target reads that would add noise and possibly make GC correction less efficient.

- looking at high coverage data, it seems there is some other bias on top of the GC content that we don't quite understand where it is coming from. It still seem somehow related to GC, but it is not linear (sometimes higher GC means more reads, sometimes fewer) It could be metilation, or something else... Anyway performing the GC correction decrease the noise.

In general, I don't see why CNAnorm should not work with your data. As I was suggesting here maybe start with relatively large windows (--window 100000) when using, However we did not have a capture experiment in mind while designing it and there might be some unforeseen difficulty.

Best of luck, and let me know

Last edited by stefanoberri; 01-23-2012 at 01:09 PM.
stefanoberri is offline   Reply With Quote

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

All times are GMT -8. The time now is 08:11 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2021, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO