[Bioperl-l] pipeline

Elia Stupka elia@fugu-sg.org
Tue, 12 Mar 2002 14:21:19 +0800 (SGT)


> I would imagine that it would be a better to have a ensembl-pipeline
> that runs bioperl runnables, and not the other way around. And this
> would be possible with specific RunnableDBs wouldn't it?

I agree. I think we should write a base classa and a clear interface for
bioperl Run modules, so that in ensembl we can write a generic wrapper
around bioperl runnables.

Ideally I think the pipeline should eventually become a nice stand-alone
unit that can work for smaller, non genome annotation tasks that have
nothing to do with ensembl and are based around bioperl run modules, as
well as ensembl runnables and genome annotation. 

It is clear that as part of the revamping of the pipeline we need to
increase radically flexibility about where inputs come from and where
output is written.

The other issue is how to reuse pieces of schema and adaptors across the
board. The example I am faced with now is:

1)Generate protein families and store them in compara (Compara code)
2)Pull out families and generate alignments (pull out of compara db, use
bioperl Run module to run clustalw)

3)Store alignment using cigar line and reference to the sequences. Now,
here ideally I would like to develop tables and adaptors in bioperl-db so
that people can store and fetch alignments when using bioperl. On the
other hand I want to be able to use these tables and adaptors within
compara, because that is where the protein sequences are stored.

I have a novel idea, but I am not sure what people think about it. Can we
make units of schema+adaptors reusable modules? i.e. put the sql in my
compara schema, and then in a meta table register that those tables are
under the control of the bioperl-db adaptors.

This way I keep development in bioperl-db, but I get to reuse the code
without making it part of compara.

What do people think about it?

Elia

-- 
********************************
* http://www.fugu-sg.org/~elia *
* tel:    +65 874 1467         *
* mobile: +65 90307613         *
* fax:    +65 777 0402         *
********************************