Entering edit mode
9.5 years ago
suzanne.mcgaugh
▴
60
Hi,
I have a VCF file and realized that one individual is an outlier and is often different than the others. I'd like to remove this individual and have the ALT and REF columns follow suit. For example, if only this removed individual had the ALT allele the ALT column would then have a "." in it instead of the ALT allele that is no longer present in the remaining samples. And same with indels.
I have tried this with VCFtools and VCFlib, but both do not recode the ALT and REF columns. Does anyone know of a tool that can do this without having to remake the entire VCF file?
Thank you,
Suzanne
Thank you for your help. I am using your software and it is getting me most of the way there. Thanks!
Would you please look at these lines:
I think the ref allele and INFO/FORMAT in 173 and 175 case no longer need the indication that it was deleted because those lines no longer exist in the dataset. Is this intentional?
Thanks very much in advance!
yes, you can remove those lines with option '-r' ( remove variant if there is not any called genotype on the line. )
Thank you for your quick response, here are my commands:
I also tried:
Both give:
I apologize in advance, if I am doing something incorrect in the command line. If you would please advise, I'd very much appreciate it.
My bad: all your variants are 0/0 (homref) and not 'uncalled'. You could quickly remove those variants by piping the vcf to https://github.com/lindenb/jvarkit/wiki/VCFFilterJS
Thank you for your response. Things aren't quite working, I tried both with the
-ecommands and without. Please see below.But just to be sure to clarify, I would like to keep all sites in the file (i.e. not remove the 0/0 sites), I simply want to recode the REF, ALT, and INFO appropriately to reflect the data currently in the "new" vcf once I remove the rouge individual.
you're not using those tools the correct way. you should do something like:
Furthermore, if all your sites are 0/0 or undefined (./.) there will be no ALT alleles.
Thank you for your response. I have your code working as above. Unfortunately, this is not the solution I wanted, as it removes the invariant sites.
I still want all invariant sites included in my file. VCFtools and VCFlib leave a signature of the removed sample in the ALT and REF columns, even if the remaining individuals do not have the alternative allele and all remaining individuals are
0/0or./.These tools may still provide something other than '.' in the ALT column. I simply want to remove one individual from a vcf file, and have the ALT, REF, and INFO recoded as if I had made the original vcf file without that individual.vcfcutsamples.jar almost does this, but it leaves the original deletions relative to the reference from the individual that was removed denoted in the REF tab (as shown in my examples above, everyone is
0/0or./.but the REF allele shows the deletion in the sample that was removed). I can work with that and just write some code to deal with it downstream.I'd be curious to know if there is actually software out there that recodes everything properly, though because it would be easier overall.
Thank you for your help.