Genomic Positions off in db135.b37.vcf

[ymc]

I noticed some indels in db135.b37.vcf I downloaded from

ftp://[email protected]

For example:

19 51835892 rs11402251 T TG . PASS GENEINFO=VSIG10L:147645;GNO;RSPOS=51835893;SAO=0;SLO;SSR=0;VC=DIV;VP=050100000000000100000200;WGT=0;dbSNPBuildID=120
19 52004791 rs67024588 G GC . PASS GENEINFO=SIGLEC12:89858;GNO;RSPOS=52004794;RV;S3D;SAO=0;SLO;SSR=0;VC=DIV;VP=050300000000000100000200;WGT=0;dbSNPBuildID=130

But we can see that the correct position is in "RSPOS=" but the second field is off.

Is this a bug or a feature???

[Heisman]

I found something that I think was similar and concluded was a bug months ago; I emailed them and one of them acknowledged it was a bug. I check back with them a few weeks later and I was told that he passed the information on to the relevant person and I haven't heard back. This was at least a few months ago.

It may have been completely different then what you just posted but my conclusion is that I would not be surprised if there are numerous errors where things are off by 1 bp or maybe a couple more for indels.

[ymc]

Oh I see. I think I will just fix that file with RSPOS positions. Thanks for your reply.

[Heisman]

Oh I see. I think I will just fix that file with RSPOS positions. Thanks for your reply.

I recall considering that and realizing it wasn't that easy. Make sure you check a decent amount after changing them and confirming they are correct (again, whatever problems I saw may very well have been fixed by now).

[ymc]

I am fixing this by hand now. How should I fix rs67024588? Are the alleles also wrong because at chr19:52004794 is C?

19 52004794 rs67024588 C CC . PASS GENEINFO=SIGLEC12:89858;GNO;RSPOS=52004794;RV;S3D;SAO=0;SLO;SSR=0;VC=DIV;VP=050300000000000100000200;WGT=0;dbSNPBuildID=130

Is this ok???

[Heisman]

According to the website it's a G: http://www.ncbi.nlm.nih.gov/projects...gi?rs=67024588, so it depends on what strand you are annotating with.

[ymc]

But doesn't VCF always show forward strand??

http://www.broadinstitute.org/gsa/wiki/index.php/Understanding_the_Unified_Genotyper%27s_VCF_files

It is a G only if it is in the reverse strand. And the base before the event is a C in the forward strand, right?

[Heisman]

Ah, yes, you're probably right. My fault.