Picking random SNPs from 1000 Genomes using Vcftools

Question

[pifferdavide]

I need to pick random sets of SNPs using Vcftools from 1000 Genomes variant set files. Is there a command to do this?

[dandan]

What kind of output are you looking for? A smaller vcf with random lines from 1000 Genomes vcfs, or just a list of SNPs (rs ids, or list of chr,position,ref,alt)?

[pifferdavide]

A list of SNPs (rs ids)

[Pierre Lindenbaum]

>-(

Answer 1

[Alex Reynolds]

To sample without replacement with sample:

$ N=1234
$ sample --sample-size=${N} foo.vcf > sample.${N}.vcf

[pifferdavide]

It looks like sample is not a Vcftools command

[Pierre Lindenbaum]

Alex clearly pointed to a tool that is not vcftools.

[pifferdavide]

Reread my question "Picking random SNPs from 1000 Genomes using Vcftools". Wrong answer since I asked how to do that job using vcftools!

[Devon Ryan]

To paraphrase the great English philosopher Mick Jagger, "You can't always get what you want. But if you ask some time, then you might find, there's a different tool that will actually do what you want."

[pifferdavide]

Sure. I tried to install your downsamplevcf, but there are too many previous steps. I installed jvarkit but it still won't work. The ant command isn't recognized. I suppose I'll have to install Apache Ant too? Sorry for these newbie questions...

[Alex Reynolds]

My advice would be not to use this answer.

[seta]

Sorry, which answer do you mean? please kindly let me know if you have any suggestions.

Answer 2

[Pierre Lindenbaum]

I wrote a simple tool to downsample vcf files: https://github.com/lindenb/jvarkit/wiki/DownSampleVcf

$ curl -skL "ftp://ftp-trace.ncbi.nih.gov/1000genomes/ftp/release/20130502/ALL.wgs.phase3_shapeit2_mvncall_integrated_v5a.20130502.sites.vcf.gz" |\
gunzip -c | java -jar downsamplevcf.jar -n 100 > out.vcf

[pifferdavide]

I installed jvarkit but it won't let me install downsamplevcf. I get the following error message. Curl command not found.

How do I install curl?

[Pierre Lindenbaum]

http://stackoverflow.com/questions/3647569/how-do-i-install-curl-on-cygwin

[pifferdavide]

Awesome. But now it doesn't read the command "ant". What do I need to install next?

[seta]

Hi Pierre,

Using downsamplevcf.jar, is there any possibility to get random SNP with similar LD and allele frequency to the our SNPs under study?

Answer 3

[Adam]

There's no simple way of doing this directly in vcftools (although using 'sample' seems a good suggestion). However, perhaps you could use the --thin command to achieve what you need?

[pifferdavide]

Never mind, I found SNPSNAP