add undefined annotations from vcf files

Question

[bioguy24]

Is there a way to find and add all undefined annotations in a vcf?

For example, if I have 3 vcf files in a directory, find all undefined annotations in each vcf, write them to a header file, then add that header file to the original vcf.

I guess something like bcftools reheader, but that seems to only work on individual vcf files.

 # find undefined annotations
 bcftools view -h file1.vcf > file1_header.txt

 # edit original vcf
 bcftools reheader -h file1_header file1.vcf > file1_fixed.vcf

then loop through the directory doing the same for file2 and file3. I have been trying to figure this out for a few weeks and not able to do so, is there a better way? I would then use bgzip and tabix on the fixed.vcf to further process. Thank you :).

[bioguy24]

Thank you very much :).

Answer 1

[Pierre Lindenbaum]

 # find all annotations

grep -E "^##(INFO|FORMAT|INFO)=" file1.vcf | sort | uniq  > file1_header.txt


#add those annotations to the other vcf files. Hoping there is no synonymous ID (same header ID, not the same meaning)

for F in other*.vcf ;
do
#find annotations in F   
grep -E "^##(INFO|FORMAT|FILTER)="  ${F} | sort | uniq > tmp1.txt

#get missing annotations

comm -23 file1_header.txt  tmp1.txt > tmp2.txt

# add those missing annotation before the #CHROM line

grep  "^##" ${F} > ${F}.new.vcf
cat  tmp2.txt >> ${F}.new.vcf
grep -v "^##" ${F} >> ${F}.new.vcf

rm tmp2.txt
done

rm  file1_header.txt

[bioguy24]

Thank you very much for your help :).