Thank you very much! This has been helpful.
I first apologise if these are obvious questions to many, but it's unclear to me the following:
- The process of variant calling does not necessarily retrieve all variants in the genome? I am asking this because I have tested some sample data from WES. I used the VCFs files provided in a website. I am interested in only a portion of all the genome's variants (200k). When I filter those by rsid in the WES VCFs, only very few are retrieved (15k).
- Does the number of retrieved variants depend on the coverage? That means, the higher the coverage, the higher chance I'll get the variants I am interested in sequenced.
- I am also thinking that there could be naming issues between the rsid in the VCFs and my rsids?. Is therefore better to subset the VCFs based on positions?
Thanks a lot for your time.
1 answer
The process of variant calling does not necessarily retrieve all variants in the genome?
it depends of the arguments of your caller, but the most frequent answer is NO
When I filter those by rsid in the WES VCFs, only very few are retrieved
you won't find them if your rs## variant is rare
Does the number of retrieved variants depend on the coverage?
it shouldn't but the higher is the coverage, the higher will be to find the difference between the real variation and the background noise in the reads.
I am also thinking that there could be naming issues between the rsid in the VCFs and my rsids
you shouldn't use the rs# id as a unique identifier for your variants. rs## can be deprecated, badly associated, etc... https://github.com/broadinstitute/gnomad-browser/blob/0f6623739f5b8a090dded09a4a151b178220c49a/browser/src/VariantPageRouter.js#L82
Log in to answer this question.