Re: [GUSDEV] Sugestion for the ISF plugin

[Steve F. <sfi...@pc...>]

the problem with deploying the increased error threshold as a solution 
in this case is that you have just pushed the problem to the downstream 
processes. 

steve

Ricardo Balbi wrote:

> Ok Chris,
>
>   We'll see this.
>
> thanks,
> Ricardo
>
> On 3/7/06, *chris mungall* <cj...@fr... 
> <mailto:cj...@fr...>> wrote:
>
>
>     You can increase the threshold at which the severity of a problem
>     causes the unflattener to throw an error.
>
>     $unf->error_threshold(2);  # do not throw for problems of severity <2
>
>     see the docs for more details
>
>     The policy here is up to you. As we discovered the last time this came
>     up, this particular genbank record has an inconsistency. It's up
>     to you
>     whether you want to load the exons from this file, or if you would
>     prefer to reject the entire record
>
>
>
>     On Mar 7, 2006, at 10:36 AM, Steve Fischer wrote:
>
>     > in this case, i am not sure this is possible.  the problem is during
>     > the bioperl parsing.   if you would like to do the work to
>     investigate
>     > whether the bioperl parser (and the unflattener) have options for
>     > this, that would be a first step.
>     >
>     > steve
>     >
>     > Ricardo Balbi wrote:
>     >
>     >> Hi all,
>     >>
>     >>   Ahead of all the problems that we are facing to load a
>     GenBank file
>     >> using the InsertSequenceFeatures (we thougth in coming back to use
>     >> the GBParser), we have one sugestion:
>     >>   Maybe modify the ISF to separate the GIs that present problems in
>     >> separate files, as GBParser does, not interrupting the
>     execution.  We
>     >> believe that this could help solve many of the problems that we are
>     >> finding (and I believe that other groups too) and would very
>     easy to
>     >> correct these problems without interrupting the work that we are
>     >> carrying at the moment.   We believe that with this
>     modification, the
>     >> ISF would become much more flexible and easy to work.
>     >>
>     >> thanks in advance,
>     >> Ricardo
>     >
>
>