This site is now read-only. You can find our new documentation site and support forum for posting questions here.
Be sure to read our welcome blog!
SV discovery batching
we've recently seen in a project of 96 samples (all ~30x) that we get a very broad range of number of calls per sample (~600-900 deletions from SVdiscovery, using only PASS calls and removing redundancies with 80% reciprocal overlap). If we only use the 30 most extreme samples (15 with highest, 15 with lowest number of calls), we get a much more homogeneous picture of 700-800 deletions per sample.
It appears that deletions that are confidently called in the batch of 30 samples are missing from the batch of 96 samples. To me this seems quite counter-intuitive, I would expect the number of SVs to increase when adding samples.
Looking into the VCF file, I see that some calls are filtered out through the DEPTH or COHERENCE filters, in some cases calls are PASS but the sample is not contained in the GSSAMPLES field, in other cases the SV is apparently not called at all. The second reason (not in GSSAMPLES) seems to be the most common one. Do you have an idea why that might be happening?
In general, given 30x WGS data, how many samples do you suggest to batch together?
Thanks a lot,