[Bioperl-pipeline] runnable parameters

[Elia Stupka]

> Right now, DataMonger isn't tied to analysis. ok so if we are spawning
> a DataMonger job for mapping, then it will have to know what the
> inputs of the previous job. 

Ah? That should be easy... wasn't that handled ages ages ago by
"copy_inputs"? Oh sorry see what you mean, you need the real output of
genewise (genes) and the input that it had... so you have weird mix....

I think in this light your solution is neater.

> I figured passing parameters to runnables aren't a isolated affair. At
> the same time, its not insane for Genewise runnable to do coordinate
> mapping. 

No I agree with you, go ahead with your idea, sorry for the "intrusion"

> But then again mapping maybe something worth generalizing into a data
> monger. 

No it is already generalised in CoordinateMapper, so you are already
reusing a reusable component. And of course you are doing mapping only if
those arguments are being passed right? So it's clean for those who
"just" want to run genewise, which anyway in reality never happens.

Just another separate point. If we want to create a miniseq around the
blast hit (which is the case for human and could be the case for other
organisms) where would that be handled? I guess it's just an option in the
InputCreate you have already, so instead of min_start to max_end, it would
go over the features, and create the miniseq right? But what about
"de-mini-seq'ing" in your design where runnable has parameters, would you
pass in a long cigar-style string, i.e. an array of start-ends?

We need to keep it in mind so we don't get burned later...

Elia

********************************
* http://www.fugu-sg.org/~elia *
* tel:    +65 6874 1467        *
* mobile: +65 9030 7613        *
* fax:    +65 6779 1117        *
********************************