[ensembl-dev] Effects predictor version 2

[Will McLaren]

Hi Stuart,

On 17 May 2011 12:04, Stuart Meacham <sm766 at cam.ac.uk> wrote:
> Hi,
>
> I am trying to use the updated version of the Effects predictor (version 2).
> I have it running now but I came across a couple of issues:
>
> I found it impossible to pass DB parameters to the script to define, for
> example, a local version of the latest human core DB. Looking at the code it
> seems that the script is expecting a 'host' parameter yet the README states
> a 'db_host' parameter (the 'd' option does not work either). Despite
> noticing this I still couldn't get it to work via command line options. In
> fact I had a lot of trouble passing any short-form options, instead I had to
> define a registry file, which worked fine.

Apologies, this is a mistake in the README. The correct parameter set
is --host, --user, --port and --password. I changed this from v1.0 to
v2.0 but obviously neglected to update the README.

I also sometimes find I can only connect with a registry file; I think
this is an issue with the Registry module itself and the way it works,
I will look into it. I suspect there's something configured on e.g.
our ensembldb.ensembl.org that allows you to connect using
$reg->load_registry_from_db().

>
> On running I get a warning about version numbers however I have only just
> downloaded and installed the latest API libs and created the core DB today.
> The warning is strange anyhow in that it states my DB version is 62 and my
> software version is 63 (which I didn't think had been released yet!).

This is strange - are you sure you are checking out the branch and not
the head of the API? You should be doing something like:

cvs checkout -r branch-ensembl-62 ensembl
cvs checkout -r branch-ensembl-62 ensembl-variation

>
> The script silently over-writes an existing output file of the same name,
> this seems a bit brutal, perhaps the default should be to fail if the file
> exists.

I think this is pretty standard behaviour for command-line programs. I
could change it to only run if in an output file name is specified
perhaps?

>
> I don't seem to get any Polyphen or Sift predictions. My command is:
>
> perl ./variant_effect_predictor_2.pl -r reg.pl -i ./variants.txt.vcf
> --format vcf -w -b 100000 --sift=p --polyphen=p --failed=0 -terms=so
>
> and after analysing chr 1 with 360,000 variants I don't see any predictions.
> Maybe this is to be expected, I'm a bit clueless about this I'm afraid.

That's also odd - any variants classified as non-synonymous coding
should have a "SIFT=*" entry in the final column. Can you try the
attached file as input on your system?

Cheers,

Will

>
> Cheers
>
> Stuart
>
>
>
>
> _______________________________________________
> 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/
>
-------------- next part --------------
7 114302130 114302130 C/A 1