[ensembl-dev] Variant effect predictor

[Henrikki Almusa]

On 2011-06-15 15:32, Will McLaren wrote:
> Hi Henrikki,
>
> On 15 June 2011 12:05, Henrikki Almusa<henrikki.almusa at helsinki.fi>  wrote:
>> Hi,
>>
>> I'm testing the variant effect predictor (2.0 against local copy of ensembl
>> 62) and noticed some things.
>>
>> 1. When I run one sample' indels with it, the resulting list has lines which
>> start with '+' instead of the uploaded variation information. What causes
>> this?
>
> I'm not sure what might be causing this - can you send me a sample
> line of input that causes this problem?

I managed to track this down and actually it was caused by an error in 
my own file transformation (I had comma instead of forward slash in 
allele). I had a situation where there was two variant alleles and my 
script didn't convert the comma between them. Sorry for the noise on 
this one.

>>
>> 2. It seems that the whole_genome option switches off the gene name. Is
>> there an option to switch it back on? I could use HGNC to get that name, but
>> would still lack the ensembl identifier.
>
> At the moment this option is missing - it will be added in the next
> release of the script.

Is there any plan when the update happen (so that I should go to 
download a new version)? There seems to be somewhat a difference in 
speed when putting exome sample though variant effect predictor between 
using the whole_genome and not.

> The reason the gene name is switched off is that the script has to
> query the database for every transcript to get the gene name, which
> greatly slows down the process.
>
> It is also trivial to get the gene ID from the transcript ID - either
> search for the ID on the Ensembl website, or query for it using the
> API ($gene_adaptor->fetch_by_transcript_stable_id('ENST000000001') )
>
>>
>> 3. The variant effect predictor seems to be able to read in the ucsc used
>> 'chr1' type chromosomes. However it does stumble with 'chrM' which is 'MT'
>> in ensembl.
>
> Thanks for spotting this - it's a bug, we should convert chrM to MT.
>
> Regards
>
> Will McLaren
> Ensembl Variation
>
>>
>> --
>> Henrikki Almusa
>>
>> _______________________________________________
>> Dev mailing list    Dev at ensembl.org
>> List admin (including subscribe/unsubscribe):
>> http://lists.ensembl.org/mailman/listinfo/dev
>> Ensembl Blog: http://www.ensembl.info/
>>
>


-- 
Henrikki Almusa