[ensembl-dev] incorporation of dbSNP 132 into ensembl

[Pablo Marin-Garcia]

On Thu, 9 Dec 2010, Andrea Edwards wrote:

> Dear all
>
>
> I was hoping you might be able to help me with something to do with the 
> incorporation of dbSNP 132 into ensembl
>
> I was looking at rs43469576 a few weeks ago  in dbSNP and I am 99% positive 
> it said on the webpage in big red letters that this snp mapped to 3 genome 
> locations. I checked what ensembl does about this and i have code showing how 
> ensembl too maps it to 3 genome locations
>
>   * chr7(65859978), chr7(91674598), chr6(86134792)

?? which version of ensembl are you using?

* from mysql latest ensembl

select vf.variation_id, vf.variation_name, vf.map_weight, sr.name as chr, 
vf.seq_region_start as start  from variation_feature vf, seq_region sr where 
vf.seq_region_id=sr.seq_region_id and  vf.variation_name= "rs43469576";

[none]

* in the web
http://www.ensembl.org/Multi/Search/Results?species=all;idx=;q=rs43469576
=> only one hit for Bos taurus

   -Pablo

>
> Now when i look at dbSNP it says this variation isn't mapped to any assembly. 
> If i search dbsnp for one of those genome locations (e.g. chr7(65859978)) 
> rs43469576 is no longer in the results
>
> Now i appreciate that I am a fallible human being and could have got confused 
> but it looks to me as if dbSNP has changed and now doesn't seem to show snps 
> as mapping to multiple locations. I looked on their build release notes and 
> schema revisions etc but there is no mention of this. So perhaps i am 
> mistaken. It often happens :)
>
> But regardless of whether i am mistaken or not there still exists a 
> difference between the info in ensembl and dbsnp and I would like to know 
> what to expect from ensembl in the january release. Will i still see a 
> variation like rs43469576 mapped to multiple genome locations or will it be 
> mapped to no assembly.
>
> Which then leads to the next question, what will you get for a variation 
> feature start/end/chromosome if it isn't mapped. As i said at present ensembl 
> seems to map the rs entries in dbSNP that are unmapped so I've not come 
> across it.
>
> thanks a lot
>
>


-----

   Pablo Marin-Garcia