This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Finding Unique values on specific INFO field of the VCF file (dbNSFP, vep annotated multisample VCF)

Hello everyone! I searched the forum but coundn`t find a question that is like mine

I have a multisample annotated VCF file (with dbNSFP plugin) to which I have filtered using VEP, like so:

/scratch/ensembl-vep-109/ensembl-vep/filter_vep --force_overwrite --input_file {1} --output_file /home/filtering/2/2_{1/.}.vcf --only_matched --filter "(clinvar_clnsig is Pathogenic) or (clinvar_clnsig is Likely_pathogenic)

After filtering, the vcf files are reduced.

Is there anyway of printing the unique values of this INFO field (clinvar_clnsig)?

I tried something like

bcftools query -f "%clinvar_clnsig\n" | sort | uniq > unique.txt

But this not seem to be working, I`m trying to figure out because this is giving me: Error:

no such tag defined in the VCF header: INFO/clinvar_clnsig

But the header contains:

##clinvar_clnsig=(from dbNSFP4.4a_grch38) clinical significance by clinvar Possible values: Benign, Likely_benign, Likely_pathogenic, Pathogenic, drug_response, histocompatibility. A negative score means the score is for the ref allele

And maybe there is a more efficient way of confirming that VEP filtering did work as intended?

bcftools filter vep vcf
• 2,111 views • link
written by avelarbio46 • avelarbio46

unique is not a linux command. You want uniq.

• 0 views • link
updated by Ram Ram • written by Pierre Lindenbaum Pierre Lindenbaum
  1. Read the manual. Seriously.
  2. query -l lists samples. You want query -f
  3. You have not specific the VCF file in the command, I'm guessing that's because the command is meant to be pseudocode.
  4. You can sort -u instead of sort | uniq.
• 0 views • link
written by Ram Ram

My mistake. I used query -f but wrote wrongly in the pseudocode above. I will try to use sort -u and also specify the output of my code in the question

• 0 views • link
written by avelarbio46 • avelarbio46

1 answer

Because my file was vep annotated, bcftools works a little different and need +slip-vep option

I ended up using:

bcftools +split-vep -f "%clinvar_clnsig\n" input.vcf | tr "," "\n" | sort -u > unique.txt

+split-vep dissociates the info column, -f works just like query -f and tr "," "\n" separates the values from the query to new lines, which are the sorted and only unique values are shown. I`m not sure if this is the most efficient way of veryfing that the filtering works, but it spits out a correct answer.

• 0 views • link
written by avelarbio46 • avelarbio46

Log in to answer this question.

More posts like this