Need help with CNV..

[shyam_la]

Hi,

I'm relatively new to bioinformatics. Have mastered SNVs and indels for the most part.. Now want to try my hand with CNVs.. I'm working with tumor-normal exome data..

My question is, is it still possible to do CNV when my tumor and normal samples are sequenced to varying depths. Lets say the coverage on my tumor data is 90+ on average and on normal reads is 45+ on average. Is CNV still possible to be done? If so which tool should be used and what parameter should be altered?

The analysis need not necessarily be paired. I just want to know which tumor samples have amplifications/deletions and in which regions/genes. I'm also interested in LoH in the tumor sample.

Any help would be appreciated..

[jujubix]

I briefly did some work determining CNV events in matched tumour/normal pairs, from my experience, the difference in coverage between libraries can be corrected for by normalizing the read counts.

Some tools for determining CNV and LoH that worked for me include:

HMMcopy for normalizing, identifying and classifying CNV events in matched and/or unmatched tumour samples:

http://compbio.bccrc.ca/software/hmmcopy/

Apolloh, the accompanying software used to determine LoH events:

http://compbio.bccrc.ca/software/apolloh/

Both were used to determing CNV and LOH events in a recent study here:

http://www.nature.com/nature/journal/v486/n7403/full/nature10933.html

[shyam_la]

Thank you. They seem to be fairly complicated to use.. Will give them a shot..

I briefly did some work determining CNV events in matched tumour/normal pairs, from my experience, the difference in coverage between libraries can be corrected for by normalizing the read counts.

Some tools for determining CNV and LoH that worked for me include:

HMMcopy for normalizing, identifying and classifying CNV events in matched and/or unmatched tumour samples:

http://compbio.bccrc.ca/software/hmmcopy/

Apolloh, the accompanying software used to determine LoH events:

http://compbio.bccrc.ca/software/apolloh/

Both were used to determing CNV and LOH events in a recent study here:
http://www.nature.com/nature/journal...ture10933.html