Using bcftools isec without stripping meta data and outputting as a VCF

Question

[kshaffman]

Hi, I am trying to use "bcftools isec" for some data analysis. I have two cell lines and I am trying to find the unique mutations in each when compared to the human reference genome. The command I have been using is as follows

bcftools isec  "a.vcf.gz" "b.vcf.gz" -C -p dir

This works well, but it stripped all the mutation annotations I had. Allele frequencies, genes, etc. It also outputs the unique mutations as a text tile.

1. Is there a way to get use bcftools isec w/o stripping annotations?
2. Is there a way to get bcftools isec to output the new file as a VCF file? I would still like to be able to use my data with visualization software such as IGV.
3. If either of these does not work, is there some way I can turn a text file into a VCF file? AND is there any way to add mutaion annotations to a text file instead of a VCF file.

Thanks for the help!

Answer 1

[Giacomo Rocchetti]

Well, that's strange. This command should return you the dir folder containing the following files:

1. README.txt: it simply describes the command executed and the produced files;
2. sites.txt: A tab-separated file containing all the filtered variants (I guess this is your output);
3. 0000.vcf: A VCF file containing all the filtered variants, which in your case will be all the variants unique to a.vcf.gz.

I tried the command myself and it seems to work as expected, outputting a VCF file, annotations included. I'm using bcftools 1.12.

With that said, I'd like to know:

1. Which version of bcftools are you using?
2. Try to add the option "-O v" to the command, which specifies that the output must be in VCF format (see the documentation) and see if it solves your problem.

[kshaffman]

Thank you this solved my problem.