[ensembl-dev] variant effect prediction non_synonymous without alleles?

[mailsvl at fastmail.fm]

Hi Graham,

Is there (if you got into it already) a specific reason why some STOP
codon are not present, while many others are (and thus show a
STOP_LOST effect)?

It seems to be pretty common, some more:

10_3181003_A/G_185Q_NON_SYNONYMOUS_CODING
10_118397884_A/G_357W_NON_SYNONYMOUS_CODING
1_1887019_A/G_763Q_NON_SYNONYMOUS_CODING
1_11906068_A/G_152R_NON_SYNONYMOUS_CODING
13_47408972_T/C_472W_NON_SYNONYMOUS_CODING
16_71061495_A/G_1018Q_NON_SYNONYMOUS_CODING
19_56499279_G/C_1049Y_NON_SYNONYMOUS_CODING
20_60963055_T/C_91Q_NON_SYNONYMOUS_CODING
21_40552307_A/G_64Q_NON_SYNONYMOUS_CODING
21_9908331_C/G_130Y_NON_SYNONYMOUS_CODING
2_172180771_A/G_379R_NON_SYNONYMOUS_CODING
4_22820504_A/T_456Y_NON_SYNONYMOUS_CODING
4_152212603_T/G_329E_NON_SYNONYMOUS_CODING
8_7718187_T/G_104E_NON_SYNONYMOUS_CODING
9_27378301_A/G_145R_NON_SYNONYMOUS_CODING

And..any idea when the new pipeline will be released? :D

Thanks,
Stef


----- Original message -----
From: "Graham Ritchie" <grsr at ebi.ac.uk>
To: mailsvl at fastmail.fm
Cc: "Ensembl DevList" <dev at ensembl.org> Date: Thu, 23 Dec 2010 16:11:46
    +0000 Subject: Re: [ensembl-dev] variant effect prediction
    non_synonymous without alleles?

Hi Stef,

It looks like you have found a bug in our consequence calculation code!
The peptide allele string in this case should be */Q and the
consequence should be STOP_LOST (our as yet unreleased new consequence
pipeline gets this right). We will endeavour to fix this as soon as we
can. For the time being I think you should be able to treat all such
results (where there is no reference peptide in the peptide allele
string) as STOP_LOST events.

Sorry about this.

Cheers,

Graham


On 23 Dec 2010, at 15:25, mailsvl at fastmail.fm wrote:

> Hi,
>
> When I use the variant effect predictor (either online or with api)
> for this snv:
>
> 10_3181003_A/G 10:3181003 ENSG00000107959 ENST00000455371
> NON_SYNONYMOUS_CODING 553 185 /Q rs2279217
>
> You can see it predicts NON_SYN with only one allele in the change (
> /Q), how is this possible?
>
> Thanks, Stef
__________________
http://www.fastmail.fm