Best Cnv Software?

Question

[Adrian Pelin]

Hello,

I know there are many reviews out there, but I can't seem to find exactly what I like.

I found this software called mrCaNaVaR, and it's great since it has it's own aligner which is supposed to be designed specifically for CNV detection. What's also great about it is that it gives .bed annotation in which it tried to determine copy number of 1kb windows regions.

What I don't like about this software is that I cannot validate the results with another aligner. If I use bwa to align the reads and then view the .bed annotations from mrCaNaVaR, I find there are regions where mrCaNaVaR says the copy number is 0 meaning no coverage, but bwa find reads that cover that position quite well.

Right now mrCaNaVaR works by quantifying copy number of windows of 1kb. It would be great if it could quantify the copy number of predefined regions such as genes for example.

Does anyone know any other software similar to mrCaNaVaR that works with NGS PE data?

Adrian

[Chris Miller]

Paired tumor/normal samples or single samples? (somatic or germline CNVs?)

[Adrian Pelin]

Paired-end samples. Somatic, but not human, I work on fungal pathogens, so I am sequencing spore populations.

Answer 1

[Charles Warden]

CoNIFER is a good read-depth tool for exome or targeted exon DNA-Seq (and you can also export the normalized read depths for analysis in DNAcopy, which I sometimes find useful).

I haven't had a chance to do copy number calls on whole genome sequencing data, so I can't help provide a suggestion there (although CoNIFER will still technically work - you will just be ignoring whatever regions you don't specify as exons).

[Adrian Pelin]

Yes but even if it works, on paper it will sound weird using exam targeted software.

[Kizuna]

Hi Charles,

I have tried to use conifer to detect CNVs in targeted exome sequencing by modifying the probes.txt file. Unfortunately, the modified probes.txt file was not read by conifer once I deleted the old target regions and then added mine.. I was sure that it is a bed format and that the four columns are OK//

May you tell me plz how did you do to make conifer works on Targeted NGS ?

Kind regards,

Kiz

[Charles Warden]

I think the best solution might be to send a copy of your error messages to the developers.

http://conifer.sourceforge.net/faq.html

I'm also not sure about what you mean by "deleted the old reads added mine". I'm assuming you specified a different set of sorted, indexed .bam files for defining RPKM values (which are then used to make the copy number calls)?

I also know that you need to have more exons per chromosome than samples, but I would doubt that is an issue if you have exome data (it only has sometimes been an issue for me when working with a panel of ~600 genes).

[Kizuna]

Thanks for the link,

In fact, I have change the probes.txt file by replacing the default baits with the ones that I have in my targeted sequencing.

In total, I was analysing 121 genes, consisting of 1,098 targets regions..