CNV between twins...

[milesgr]

I am looking to use a CNV tool to find CNV between twins (not just between a single sequence and a reference genome. I am attempting to use CNVnator, which is supposed to be the most accurate CNV detection application. Is this the right approach? I know CNV-Seq is another approach (and allows you to use one sequence as test and one as reference, but the documentation is limited and I'm not certain that by following the example in the manual that I am obtaining exactly the results I want (and with certain outputs, I'm not even sure what I am looking at). Any ideas would be appreciated, especially if they pertain to CNVnator. Thanks.

[husamia]

I am also looking for CNV detection in twin sequencing whole exom data. I found this project http://code.google.com/p/condr/

[milesgr]

That looks very bare-bones but seems workable as long as you know what you're doing. Have you had any success running this? Do you think it would work for whole-genome sequencing input? Also, what kind of input does it take? I have aligned BAM.

[husamia]

Originally Posted by milesgr

That looks very bare-bones but seems workable as long as you know what you're doing. Have you had any success running this? Do you think it would work for whole-genome sequencing input? Also, what kind of input does it take? I have aligned BAM.

I have illumina whole exom sequencing with nimblegen exom v2 enrichment. Paired end about 65bps. What type of data do you have? I don't have BAM alignment but I will use galaxy to get it. I havn't tried this code yet. I am going it if i find it appropriate. Did you get any results with CNVnator?

[milesgr]

I have data from whole genome sequencing done using 90 bps paired-end reads. I have aligned it using bowtie with parameters specified by the authors of CNVer, who have suggested a CNV-focused alignment. The params are -v 2 -a -m 600 --best --strata

Upon converting from SAM to BAM after alignment, I installed root (that was quite the chore but somehow it worked). I then ran CNVnator using two bam files as the "tree" inputs, each one representing one set of the paired ends. Unfortunately I received an error that I am unable to decipher. I will post it here in a minute in a separate post. I'm not sure what I am doing wrong...I e-mailed the author, so hopefully I can figure out the issue. Apparently SegSeq is quite a good CNV tool, but very difficult to implement. If you or anyone else have success with something that would be useful for twin CNV comparisons, please share (I will do the same). Thanks.

[milesgr]

Here is the error:

Assuming BAM file ...
terminate called after throwing an instance of 'std::bad_alloc'
what(): St9bad_alloc
/campus/apps/amd64_linux26/packages/torque-2.4.8/var/mom_priv/jobs/7319.localhost.localdomain.SC:
line 10: 30301 Aborted ./cnvnator -root
/output_folder/BAM_output.root -tree /input_folder/input_file1.bam
/input_folder/input_file2.bam

[husamia]

Originally Posted by milesgr

I have data from whole genome sequencing done using 90 bps paired-end reads. I have aligned it using bowtie with parameters specified by the authors of CNVer, who have suggested a CNV-focused alignment. The params are -v 2 -a -m 600 --best --strata

Upon converting from SAM to BAM after alignment, I installed root (that was quite the chore but somehow it worked). I then ran CNVnator using two bam files as the "tree" inputs, each one representing one set of the paired ends. Unfortunately I received an error that I am unable to decipher. I will post it here in a minute in a separate post. I'm not sure what I am doing wrong...I e-mailed the author, so hopefully I can figure out the issue. Apparently SegSeq is quite a good CNV tool, but very difficult to implement. If you or anyone else have success with something that would be useful for twin CNV comparisons, please share (I will do the same). Thanks.

So you have samples sequencing data from twin? you interested in finding shared cnv or nonconcordant cnv? can you clarify?

[milesgr]

I am interested in finding both, but the most important ones we want are nonconcordant CNV's. One twin has cancer, the other does not. We want to know CNV's responsible for this. It also would help to call concordant ones simply for validation however.

[milesgr]

Has anyone had any success with FREEC? This seems to have the capability to do the trick.

[husamia]

Originally Posted by milesgr

I am interested in finding both, but the most important ones we want are nonconcordant CNV's. One twin has cancer, the other does not. We want to know CNV's responsible for this. It also would help to call concordant ones simply for validation however.

do you know if they are monozygotic or dizygotic twin? either case they would be very similar genotypically. Theoritically how much of their exom would be concordant based on knowing the'r twin? taking into account sequencing noise? how much background noise do you think there is generally in exom sequencing?