%varfreq homozygote off to 30%

[Zaag]

We do resequencing on the 454 and the SOLiD. We do whole exomes (Nimblegen) and we do a small custom region (Nimblegen, 1kb).

When we plot a histogram of our variant frequencies we get what we expect for the small region on the 454 but when we do the same for the whole exome or small region capture on the SOLiD there seem to be not enough reads with variant and when we try'd to validate with Sanger we got homozygotes in Sanger that were off to 30% variantfrequency on the SOLiD capture

Does anyone have a thought on this?

[Richard Finney]

I'm not a wet lab kind of guy, I deal more with big data ... but ... check to see if there's a known segment duplication/deletion/variation in your region of interest. The UCSC browser has these "segmental variation" tracks. You might need to adjust the region that you're cutting out based on this information. A good example of a "cryptic duplication" is at the PRIM2 gene area mapped against the reference (higher whole genome coverage there and tri-nucleotide SNPS).