Ideas how to annotate vcf with local variant freq info?

[swNGS]

I'm hoping for some inspiration on something that I suspect is not too tricky, but I would like some pointers.

I have a targeted sequencing panel for a particular phenotypic group of disorders in humans, which is run on a regular basis by our group. For each sample, we then filter a functionally annotated variant dataset (in excel, it's a small amount of variants numbering in the 100's per sample). *I'm using Annovar for the annotation part.

What I want to do is annotate either the original vcf per sample, or the Annovar annotated datasets variant list with the frequency that a given variant has been identified by our group FROM ALL PREVIOUSLY SEQUENCED SAMPLES. I would envisage that this could be achieved by either holding all variants locally in some sort of database structure and somehow using that to annotate the pre-annovar vcf, or somehow recursively loop through all vcfs for previous samples to generate allele frequency info and then use this to annotate the pre-Annovar vcf.

Any ideas would be appreciated !

Thanks

[jhersheson]

I am trying to do something similar with our exome files to filter out sequencing artifacts. It should be pretty easy with your target seq data.
What I would do is generate a unique signature for each variant - something like Chr_Star_Ref_Obs in the annovar annotated file. Then create a concatenated list of all of these signatures and use awk (or excel) to report the count of each variant in that list and hence the frequency. You can then just keep the unique signatures and use that list to annotate the annovar files with the variant freq in your dataset.

If anyone has any suggestions for doing something similar with exome files I'm all ears!

[swNGS]

That sounds promising, but I think the more long term option is to database all variant results, at this stage I think this would have to be just the bare necessities in the vcf, ie chr pos ref alt +sampleID, so that in addition to potentially being able to annotate a 'new to the assay' vcf with an additional column of something that would say eg 2/150 (seen 2x out of 150 samples), I could then potentially figure out which other samples it had featured in.

I'm inclined to not work with annotated data to store variant frequencies since the format/content/application used to annotated will/may change over time.

I think the solution could be the same for exome since its essentially just another flavour of targeted sequencing, albeit a bit bigger than the the 250kb I'm dealing with !

[jhersheson]

My grasp of scripting and databases is pretty limited so I tend to go for the easiest although not always the most efficient method. If you keep one file with all of your sample data in, you should be able to just use filters in excel to show you which samples the variants are in. I guess it depends on how many variants are in each file, excel is fine for about 1million rows but you could always filter the text file in the command line if it was any bigger.

Re using annotated data, there isnt really much difference between using the annovar file and the vcf file as you are only going to be using the chr_start_ref_obs fields which are the same in both files. Annovar just uses the variant calls found within the vcf file albeit in a different format as input

[Nino]

I seem to have trouble posting a new thread so since this the most relevant place for me to post my question. I am trying to create a vcf file (obviously in vcf format) of the spop gene in human from hg19. If anyone has some insight on how to do this and how to create a thread on seqanswers please let me know.

Thanks,
Nino

[GenoMax]

If anyone has some insight on how to do this and how to create a thread on seqanswers please let me know.

Thanks,
Nino

Select "Forums" link (on left) from main SEQanswers page.
Choose the correct "Forum" you want to post in and click on the name (e.g. General).
At top left on the Forum page, choose "New Thread".