Tabix For Retrieving High Quality Snps From Sanger Mouse Project

Question

[Sakti]

Dear Biostars,

I'm trying to get all the high quality SNPs of the strains 129S1 and 129S5 from the Sanger Mouse sequencing project. For this I have used tabix previously, and specifying chromosomal ranges of my interest and then filtering based on strain and call quality (i.e. high confidence SNP). Usually I use the following command:

tabix -h ftp://ftp-mouse.sanger.ac.uk/REL-1111-SNPs/mouse-snps-all.annots.vcf.gz 7:123000000-124000000 > 7_123Mb_124Mb_129S1_5_SNPs_Sanger.txt

However doing this always gives me gigantic (>1Gb) files which I then have to process to find my 129S1 and 129S5 SNPs compared to C57Bl6 reference.

This time I want to do the whole genome and get all high confidence SNPs using tabix, but I'm confused with the syntax and how to specify the program to only retrieve S1 and S5 high confidence SNPs compared to the reference.

Anyone done something similar???

Thanks!!

Sakti

[Rm]

for whole genome dont limit to a specific region as you did it above; its better to download the vcf.gz n vcf.tbi files and run them locally

Answer 1

[Ashutosh Pandey]

For call quality you can use the ones that have been marked as "PASS". Basically they have already done the QC for you. To get the SNP calls for particular strain, you can use Genotype information (GQ: tag in VCF file) that is mentioned for every strain that they sequenced. Make sure the calls are homozygous. Read more about vcf format from http://www.1000genomes.org/wiki/Analysis/Variant%20Call%20Format/vcf-variant-call-format-version-41 here.

[Sakti]

Thanks! Yes, but my question is how to indicate the PASS on the tabix command for filtering, plus the strains, that's where I get stuck :(

Answer 2

[Ashutosh Pandey]

The tabix outputs the filestream that is in vcf format. You can use unix pipe "|" to grep "PASS" to get all the SNPs that passed the QC. You can then use vcftools (http://vcftools.sourceforge.net/) in pipe and extract what you want.

Some command like

tabix -h ftp://ftp-trace.ncbi.nih.gov/18mousestrainsgenomes/ftp/release/someXYZ.genotypes.vcf.gz 1:1000000-1001000 | perl vcf-subset -c Sample1 | bgzip -c sample1.vcf.gz

This is not the exact command but sth similar to it should work. Read vcf tools to see which command will work better for you.

[Sakti]

Thanks! Will try this one and see the output :)