[Biojava-l] Need a hand?

Matthew Pocock matthew_pocock@yahoo.co.uk
Sun, 21 Jul 2002 20:42:42 +0100

Shailesh L Mistry wrote:
>>>a) phylogenetic trees - object model for trees, some simple algorithms
>>>and IO filters
> do you mean doing the actual analysis or just reading/writing phylogenetic tree
> formats and a GUI to view them with?

We obviosly would need the IO and possibly GUI stuff. Analysis would be 
good. Perhaps the bioperl object model would be a good place to start.

> there is already quite a bit of software out there that does this and may be
> useable (with permission)

Sure. Would you like to see if there is any Java code that we could 
snaffle up? I guess for the analysis we can produce biojava interfaces 
for things that can build trees, and implement this with calls out to 
external packages or to java code depending on what people want to do 
that day.

>>>b) emboss integration - launch emboss apps either via command line or as
>>>native methods
> do you mean biojava ->emboss or biojava->jemboss->emboss?

Either one. I think you can convince emboss to generate custom stub code 
(using their program invocation meta-data stuff), so perhaps a realy 
thin customization layer (like jemboss) that sucks emboss apps in 
direct, and also auto-generating java->external_process stubs. I don't 
know what is easiest and most platform neutral.

>>>We also never got an expression data API up.
> not sure exactly what it is you need doing here.
> Please can you elaborate on this.

We have no API at all for representing expression data. There are 
emerging standards, so it may be a good idea to see if we can play well 
with these. Perhaps this is lower down the list than trees and emboss.

> Shelly.