[Cdk-devel] Re: [QSAR-devel] Molecular Open Web Services

[Peter Murray-R. <pm...@ca...>]

At 16:12 15/12/2004 +0100, E.L. Willighagen wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On Monday 13 December 2004 13:36, Matteo Floris wrote:
> > > What services would people like?
> > > Our own list includes but is not limited to:
> > > - calculate 2D coords for connection table
> > > - calculate 3D coords for molecule
> > > - MM, MD, MO calculations
> > > - calculate descriptors for molecule
> > > - calculate properties for molecules
> > > - legacy conversions
> > > - search databases
> > > - create SVG diagrams for molecules and reactions
> > > - create PNGs
>
>Peter, the system you describe, is that using SOAP?

I shall let YY answer definitively, but I am sure it does.

>  So that it could integrate
>with SciTegic's PipeLine Pilot, which can use SOAP services as nodes in the
>pipeline?

I you have PLP and want to integrate, then fine. I have worked with PLP and 
it is a good product. However I do not know its data architecture but I 
strongly suspect it is not compatible with what we are developing here.

>At the CIC Workshop I met Asutosh Yagnik of SciTegic and talked with him a 
>bit
>on their software, and talked about integrating CDK functionality with
>PipeLine pilot... And it was reasonably straightforward to integrate SOAP
>services...

My concern is primarily about the data objects, particularly the XML 
serialisation. If PLP does not support them then it may not be easy to do it.

What I am suggesting here is that we look at systems like Taverna (or 
Kepler). We have built a very successful workflow with Taverna that works 
well. There is quite a large development team and they are very responsive. 
They already have Jmol as a viewer. They have a model where (say) other 
renderers such as JCP could be integrated. The system can run as 
standalone, web services grid-enabled, etc.

The first task is for us to agree on the object we pass. These are safest 
if they are serialized (e.g. in CML) but it appears that for non-persistent 
objects Java will work (works fine for me). It could be that octet has a 
role here.

YY is currently setting up pages where you can develop your own simple WS 
aggregation. A typical example is
"find the molecular mass for a SMILES".*

We can do this in two steps:
- convert SMILES to CML (OB)
- calculate MWt (JUMBO: MoleculeTool.getMolecularMass());

The idea will be to create (very simple) workflow on-the-fly. It is likely 
that we shall be able to save these as Taverna XML objects

P.

* - I expect that some systems may be able to do this directly, but this is 
just an example. Moreover it could be that other components could be 
substituted for each step


>Egon
>
>- --
>e.w...@sc...
>PhD-student on Molecular Representation in Chemometrics
>Radboud University Nijmegen
>http://www.cac.science.ru.nl/people/egonw/
>GPG: 1024D/D6336BA6
>-----BEGIN PGP SIGNATURE-----
>Version: GnuPG v1.0.7 (SunOS)
>
>iD8DBQFBwFRbd9R8I9Yza6YRApFvAKCcOzK+9n5xiT2cPCAtvB9WqrEzZQCfYI73
>D1dZx7wtYnfNiMfRltni0us=
>=XDrF
>-----END PGP SIGNATURE-----

Peter Murray-Rust
Unilever Centre for Molecular Informatics
Chemistry Department, Cambridge University
Lensfield Road, CAMBRIDGE, CB2 1EW, UK
Tel: +44-1223-763069