[BioPython] Precompute database information

Fernando fennan at gmail.com
Tue Oct 16 18:55:54 UTC 2007


Hi Peter,

>How big would your pre-computed data be?  If its some sort of table or
>other simple data you could perhaps use a simple text file; Another idea
> for complicated objects is to use python's pickle module.

It would be big... I an dealing with pairwise terms comparisons and I want
to consider different species as well.

>How often would the pre-computed data need to be updated?  Every time
>there is a new Gene Ontology release?  It might be better have the
>module download and cache the latest version on request (rather than
>shipping an out of date dataset with Biopython).

Yes, I could do that... It would be OK in Biopython to use mysql? If so the
module could download the last GO version on request, install it and work
with that version until the users decides to update it.

On 10/16/07, Peter <biopython at maubp.freeserve.co.uk> wrote:
>
> Fernando wrote:
> > Hi everybody,
> >
> > I am thinking in including some algorithms that I work with into
> biopython.
> > My first concern is that I'm using a local image of the Gene Ontology
> > database to perform several operations. In order to avoid such database
> > accesses I could precompute the information I need and load it once the
> > module is called. How should I do it? Is there a guideline style to load
> > external variables or something like that? Any other ideas/suggestions?
>
> I think you need to go into more detail.
>
> How big would your pre-computed data be?  If its some sort of table or
> other simple data you could perhaps use a simple text file; Another idea
> for complicated objects is to use python's pickle module.
>
> How often would the pre-computed data need to be updated?  Every time
> there is a new Gene Ontology release?  It might be better have the
> module download and cache the latest version on request (rather than
> shipping an out of date dataset with Biopython).
>
> I don't think we have anything in Biopython that requires regular
> updates.  Things like genomes and sequence databases are left up to the
> user.
>
> Peter
>
>



More information about the Biopython mailing list