How to extract only transition mutations/ SNPs from VCF file

Question

[sadsam]

I have generated vcf files containing SNPs from mutant genome sequences. Since the mutagen only caused transitions to these genomes, I want to make vcf files that have only transition mutations/ SNPs. I know awk could easily do it. I have started with the following command:

 cat file.vcf | awk '$4=="G" && $5=="C"' > transitions.vcf

it generates a file with G>C transitions only. However, I need to include 3 other transitions (A>G, C>T and T>C) from the same file to the transitions.vcf file. I am not sure how to combine all awks together to get one single vcf file. Any help is highly appreciated.

[Ghani]

All this would do is extract positions where the REF allele is G and ALT is C irrespective of your subject ‘s genotype.

It will NOT give you positions where your subject has G > C mutations

Answer 1

[Pierre Lindenbaum]

using vcffilterjdk http://lindenb.github.io/jvarkit/VcfFilterJdk.html and the function isTransition of VariantContextUtils

 java -jar dist/vcffilterjdk.jar \
  -e 'return variant.isSNP() && variant.isBiallelic() && VariantContextUtils.isTransition(variant);' \
  input.vcf

and if you really want awk

awk -F '\t' '(($4 == "A" && $5 == "G") || ($4 == "G" && $5 == "A") || ($4 == "C" && $5 == "T") || ($4 == "T" && $5 == "C") || $0 ~ /^#/ )'   input.vcf

[sadsam]

Thanks a lot Pierre. With my current knowledge, I would prefer awk as there is no need to download/ install anything