I emphasize: hundreds. There is a good cutevariant tool, but it is suitable for a single file. VarAFT has a lot of bugs. VCF-Miner possibly good, but I didn't use it because of troubles with docker (permission denied).
I emphasize: hundreds. There is a good cutevariant tool, but it is suitable for a single file. VarAFT has a lot of bugs. VCF-Miner possibly good, but I didn't use it because of troubles with docker (permission denied).
You want a variant browser. Most GUIs are not designed to filter hundreds of samples and thousands of variants simultaneously. If you have no money and no engineering support, your choices are pretty limited.
The most mature open source variant browser is seqr. It's used by several medical schools. If you are having trouble with Docker then Seqr will likely be too difficult for you to install, as it requires several services to run.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
I emphasize: loops
you already used
parallel
Isolating and extracting regions in VCF file . So what is the problem ?Yes, I made a loop with grep for extraction of particular rsID. It was okay, but I need a more convinient tool, if it exists
Excel or any spreadsheet software? VCF is essentially a tab separated file. This is not advisable because then you have to recreate the vcf file after filtering. Use a dedicated tool,
vcftools
orvcfR
similar.Do you mean vcftools can recreate vcf from txt after filtering?
cross-posted: https://www.reddit.com/r/bioinformatics/comments/1addss8/what_tools_are_there_for_filtering_hundreds_of/