about dbsnp versions and the best way to filter your variant list with it...

[Fabrice ODEFREY]

Hi everyone,

we just got our first list of variants (in gff3) from our first exome analyses.
I have been looking around and found different info but couldn't find a definite answer about which version of dbsnp to use. we have used hg19 as our ref. we heard that 129 was the best because after that 1000genomes project had added rare variants that we shouldn't necessarly filter out...but 129 was based on hg18 so some positions won't compare... what would you recommand? realign with hg18? use dbsnp131?

then we want to write a perl script to compare our gff3 file and the dbnsp file and output a 3d file with only the variants present in our gff3 file... what would you use to compare? the start position of each variant?

thanks a lot for any help you could provide!!!

[adamdeluca]

I use snp131 but I only filter out snps that have frequency data indicating that it is too common to cause the disease I am studying.

Also I would recommend looking at bedtools to do the filtering. It can directly compare your gff file to a bed file downloaded from the UCSC table browser.

[Fabrice ODEFREY]

thanks a lot, I'm trying right now!