[Bioperl-l] EST database w/postgresql

[Ewan Birney]

On Thu, 3 May 2001, Charles Hauser wrote:

> Greetings,
> 
> I am new to the list, and to perl so please excuse my shortcommings.
> I need to set up a database (postgresql) to both manage and use for 
> web queries  EST data and associated blast results, tentative gene 
> and protein id's etc.  I was hoping some of you out there have done 
> just this and might offer up advice for table layout, pitfalls to 
> avoid etc. before I dive in.


First off - do contact Jason Staijch, if you haven't already, as he is
both at Duke and understands this area well and final is a cool guy...


Secondly - check out the bioperl-db cvs module from the anonymous cvs
server (instructions at http://cvs.bioperl.org/). This will

  (a) solve some of your problems

  (b) give you an idea how to structure code/table bindings


Thirdly - make sure you have really scoped the problem well with either
yourself (if you are doing it for yourself) or your "customers" ie - teh
people you are building it.

  - in this area I am a big proponent of Extreme Programming (XP) where
you would offer to your customers a number of options taking roughly 2/3
weeks to complete like:

   Parsing a BLAST result into a database and having a bare-bones web page
  (notice - no intiation of the BLAST stuff remotely)

Or

   Uploading an EST via the web into the database

Or 

   something else you can imagine taking 2/3 weeks (honest weeks, eg not
ideal weeks when noone interrupts you and your network stays up 100% of
the time)


Let the customers choose which things to work on first -- it (a) keeps you
focused on what they want (b) involves them in the decisions about what
will work first and what second - they will get less pissed off that not
everything works first time...


Fourthly - decide if you want to implement "inside" bioperl-db or not -
this depends mainly on your comfort level with bioperl, but I would be
happy to see - for example - additional pieces of schema into bioperl-db
to provide storage of BLAST results

 

Finally - Do the first iteration - go back to customers and possibly this
list - and stepwise build it up.



I'm happy to help -as I am sure is Jason - on schema and code design for
new bits of code either inside bioperl-db or completely outside.



ewan


> 
> Thanks,
> 
> Charles
> -- 
> Dr.Charles Hauser
> DCMB, LSRC Rm B328
> Duke University
> Durham NC 27708-1000
> 919.613.8159 tel
> chauser@duke.edu
> 
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l@bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
> 

-----------------------------------------------------------------
Ewan Birney. Mobile: +44 (0)7970 151230, Work: +44 1223 494420
<birney@ebi.ac.uk>. 
-----------------------------------------------------------------