Hello, all; first time poster and exome sequencer here. We're attempting to find differential base pairs in paired tumor/normal samples from a 2.1M Exome capture run on an Illumina IIGX. Our pipeline thus far is basically:
bowtie->samtools->ShortRead in bioconductor (to id variants)->annovar.
and it seems to be working well; we've correctly identified known variants in our sample, for example. However, we've also gotten back some SNP-chip data recently that seems to indicate considerable CNV between one of our samples and its control, and we're concerned that it may be affecting our results. So two questions:
1) It's been mentioned by some colleagues that GATK may do better genotyping than our pipeline due to the local realignment. We hadn't been too concerned since we figured exome sequencing was less indel-prone. However, would GATK (or some other tool) be preferable in terms of handling CNV variations vs. more local indels?
2) In general, is this a problem we should be concerned about? Is there a general approach or attitude places like Broad/Sanger have about CNV in their variant identification?
Thanks...
bowtie->samtools->ShortRead in bioconductor (to id variants)->annovar.
and it seems to be working well; we've correctly identified known variants in our sample, for example. However, we've also gotten back some SNP-chip data recently that seems to indicate considerable CNV between one of our samples and its control, and we're concerned that it may be affecting our results. So two questions:
1) It's been mentioned by some colleagues that GATK may do better genotyping than our pipeline due to the local realignment. We hadn't been too concerned since we figured exome sequencing was less indel-prone. However, would GATK (or some other tool) be preferable in terms of handling CNV variations vs. more local indels?
2) In general, is this a problem we should be concerned about? Is there a general approach or attitude places like Broad/Sanger have about CNV in their variant identification?
Thanks...