Problems with vcf to 1ksnp

[ivargr]

Hi!

I am having trouble using the vcf_to_1ksnp.py script to convert vcf-files to the 1ksnp format. For all the vcf files I've tried (e.g. ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr22.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz), the script always ends with printing "Found 0 SNPs" . I could try making a sample case that fails to show you, but before doing that I wondered whether you have a set of sample vcf/reference files that I could try out (maybe my vcf file is formated differently or I've misunderstood how to use the script)? I checked the test directory of this repo, and found a test (phasing.sh) that refers to a sm.vcf file, but this test fails since the sm.vcf file does not exist. Is it maybe not added to the repo?

Another quick question: Does the vcf to 1ksnp also support indels? If not, how do I get a 1ksnp file with indels?

In advance, thanks!

[milkschen]

Hi,

You may try python vcf_to_1ksnp.py --reference hg19.fa --vcf ALL.chr22.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf --out test.1ksnp

You can put ALL.chr22.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz in the test directory and use python make_vcf.py > sm.vcf to generate the sm.vcf file. In my tests it works properly. Please let us know if there are still no outputs.

In the test directory the sm.fa is only a small region of the chr22 genome. If you use ALL.chr22.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz as sm.vcf and the complete chr22 as sm.fa it's also runnable.

We'll look into the indel issue and keep you updated.

[ivargr]

Thanks for the quick reply!

The command you suggested with the ALL.chr22.phase3_shape[...] file actually turned out working fine (I'm not sure why it didn't before, I guess I have messed things up on my side). I am still having problem with a vcf file from the 1001 Genomes Project (Arabidopsis), but I have figured out that this is because this vcf file does not contain the vt=SNP tag, which seems to be required by the vcf_to_1ksnp.py script. I guess I will be able to fix this by modifying the vcf_to_1ksnp.py script.

It would, however, be great with a Python script for conversion that handles both SNPs and indels and that does not require the vt tag that specify variant type.

[zhixuqiu]

Hi, can vcf_to_1ksnp.py script convert vcf files with missing values to 1ksnp format?

[milkschen]

We'd recommend you use bcftools +missing2ref to preprocess your VCF with missing values.