UsefulChem web services

6 views
Skip to first unread message

DuncanHull

unread,
Aug 6, 2007, 7:17:07 PM8/6/07
to UsefulChem
Hello, I just met Jean-Claude Bradley at scifoo. He asked me for
advice on making taverna workflows and myexperiment workflows of
useful chem web services, so my question is, where and what are the
services? Are they SOAP+WSDL or REST or something else?

Rajarshi

unread,
Aug 6, 2007, 7:19:54 PM8/6/07
to UsefulChem

You can find a number of services (WSDL and usage docs) at
http://www.chembiogrid.org/wiki/index.php/Web_Service_Infrastructure#Cheminformatics_Services

They are all SOAP+WSDL services

Jean-Claude Bradley

unread,
Aug 7, 2007, 6:29:00 PM8/7/07
to usefu...@googlegroups.com
Rajarshi,
Duncan is asking specifically about the services described for the CombiUgi workflow here:
 
Starting with
Step 1 creating the virtual SMILES library
--
Jean-Claude Bradley, Ph. D.
E-Learning Coordinator for the College of Arts and Sciences
Associate Professor of Chemistry
Drexel University

http://drexel-coas-elearning.blogspot.com
http://drexel-coas-talks-mp3-podcast.blogspot.com/
http://usefulchem.blogspot.com

Rajarshi

unread,
Aug 7, 2007, 6:52:11 PM8/7/07
to UsefulChem

On Aug 7, 6:29 pm, "Jean-Claude Bradley"


<jeanclaude.brad...@gmail.com> wrote:
> Rajarshi,
> Duncan is asking specifically about the services described for the
> CombiUgi workflow here:http://usefulchem.blogspot.com/2007/06/combiugi-web-service.html

Aah. Well the CombiUgi virtual library page is basically Python CGI
that makes use of a web service to perform some cheminformatics
processing.

The web service is the Utility service (http://rguha.ath.cx:8080/cdkws/
services/Utility?wsdl) and the method of interest is the
getFragmentsWithClosure (http://cheminfo.informatics.indiana.edu/
~rguha/ws/cdkws/api/net/sf/ciccgrid/cdkws/
Utility.html#getFragmentWithClosure(java.lang.String[],
%20java.lang.String,%20int))


What this method does is to take a set of SMILES (reagents) and a
chemical group (specified as SMARTS) such as COOH or NH2 etc. It then
removes this group from the input SMILES and attaches a ring closure
symbol. The processed SMILES are then returned back to the caller.

Given this set of processed SMILES, it is a simple task to loop over
the various regaents and plug them into a template molecule resulting
in specific products.

So the point is: the virtual lib application makes use of a web
service for cheminformatics specific stuff, but is not itself a web
service

Jean-Claude Bradley

unread,
Aug 8, 2007, 10:03:01 AM8/8/07
to usefu...@googlegroups.com
Duncan,
Is that the information you needed to get the workflow started?
I added you to the CombiUgi project on MyExperiment.
Will we all need to install Taverna to run a workflow?
 
Rajarshi - you can join the beta of MyExperiment as described here
 


 
On 8/7/07, Rajarshi <rajars...@gmail.com> wrote:



On Aug 7, 6:29 pm, "Jean-Claude Bradley"
< jeanclaude.brad...@gmail.com> wrote:
> Rajarshi,
> Duncan is asking specifically about the services described for the
> CombiUgi workflow here: http://usefulchem.blogspot.com/2007/06/combiugi-web-service.html

Aah. Well the CombiUgi virtual library page is basically Python CGI
that makes use of a web service to perform some cheminformatics
processing.

The web service is the Utility service (http://rguha.ath.cx:8080/cdkws/
services/Utility?wsdl) and the method of interest is the

~rguha/ws/cdkws/api/net/sf/ciccgrid/cdkws/
Utility.html#getFragmentWithClosure(java.lang.String[],
%20java.lang.String,%20int))


What this method does is to take a set of SMILES (reagents) and a
chemical group (specified as SMARTS) such as COOH or NH2 etc. It then
removes this group from the input SMILES and attaches a ring closure
symbol. The processed SMILES are then returned back to the caller.

Given this set of processed SMILES, it is a simple task to loop over
the various regaents and plug them into a template molecule resulting
in specific products.

So the point is: the virtual lib application makes use of a web
service for cheminformatics specific stuff, but is not itself a web
service




DuncanHull

unread,
Aug 10, 2007, 7:38:58 AM8/10/07
to UsefulChem
Hello

On Aug 8, 3:03 pm, "Jean-Claude Bradley"
<jeanclaude.brad...@gmail.com> wrote:

> Is that the information you needed to get the workflow started?

I will have a play with it and get back to you

> Will we all need to install Taverna to run a workflow?

At the moment, yes. Once the workflow is created, you need Taverna
installed locally to execute it. In the next version, we will add the
capabililty to run the workflow from the browser.

Duncan

Jean-Claude Bradley

unread,
Aug 10, 2007, 9:26:16 AM8/10/07
to usefu...@googlegroups.com
Duncan,
Thats not a problem.  As long as it is fairly easy to execute a workflow and that it runs on Windows we can use Taverna.
Our current inputs are 4 txt files with SMILES lists of about 30 each.  Will those serve as the actual inputs when using Taverna?
Are file sizes a limitation?  One way around that would be to process the docking web service (when we get it) right after each molecule is generated by the first service and keep a running list of the top few hundred.  That way no large files with millions of SMILES codes would ever have to be stored.

 

DuncanHull

unread,
Aug 10, 2007, 10:31:29 AM8/10/07
to UsefulChem
Hello

On Aug 10, 2:26 pm, "Jean-Claude Bradley"
<jeanclaude.brad...@gmail.com> wrote:

> Our current inputs are 4 txt files with SMILES lists of about 30 each. Will
> those serve as the actual inputs when using Taverna?
> Are file sizes a limitation?

Size files are not a limitation but I need examples.

I've created a workflow of one service "getFragmentWithClosure" at
http://carlin.myexperiment.org/workflow/show/41

I need examples of smiles string, group and closure parameters in
order to make it work.

You can run Taverna from the command line also, without any need for
the GUI at all.

Duncan

Rajarshi

unread,
Aug 10, 2007, 10:52:36 AM8/10/07
to UsefulChem

On Aug 10, 10:31 am, DuncanHull <dullh...@yahoo.com> wrote:
> Hello


>
>
> Size files are not a limitation but I need examples.
>

> I've created a workflow of one service "getFragmentWithClosure" athttp://carlin.myexperiment.org/workflow/show/41


>
> I need examples of smiles string, group and closure parameters in
> order to make it work.

You can try the following set of SMILES

CC(C)(C)OC(=O)NC(C)C(O)=O
CC(C)(C)OC(=O)NCC(O)=O
CC(C)(C)OC(=O)NC(C(O)=O)C(C)CC
CSCCC(NC(=O)OC(C)(C)C)C(O)=O
CC(C)(C)OC(=O)NC(CO)C(O)=O
OC(=O)C(NC(=O)OC(C)(C)C)Cc2cnc1ccccc12
CC(C)(C)OC(=O)NC(CC(C)C)C(O)=O
CC(C)(C)OC(=O)NC(CC(N)=O)C(O)=O
CC(C)(C)OC(=O)NC(CS)C(O)=O
CC(C)(C)OC(=O)NC(CCC(N)=O)C(O)=O
CC(C)(C)OC(=O)NC(CO)C(C)O
CC(C)(C)OC(=O)NC(CC(O)=O)C(O)=O
Oc1ccc(CC(NC(=O)OC(C)(C)C)C(O)=O)cc1
OC=O
CC(O)=O
CCC(O)=O
CCCCC(O)=O
OC(=O)CCCCCCCCC
OC(=O)CCCCCCCC
OC(=O)CCCCCCC
CCCCCCC(O)=O
CCCCCCC(O)=O
OC(=O)CCCCCCCCCC
Cc1ccc(cc1)C(C)C(O)=O
Brc1ccc(CCC(C(O)=O)C(F)(F)F)cc1
Clc1cc(Cl)ccc1CCC(O)=O
Fc1cc(CCC(O)=O)ccc1
Brc1ccccc1CCC(O)=O
OC(=O)c1ccccc1

The group is C([Oh1])=O
The closure parameter is 90

jcb

unread,
Aug 15, 2007, 1:44:53 PM8/15/07
to UsefulChem
Duncan/Rajarshi,
I understand that you have a simple part of this workflow working now:
appending a closure tag to a SMILES tag. Could you please provide
instructions here on the mailing list so my students and other
collaborators can replicate this?
What is the command to use on Taverna?
Thanks!

Jean-Claude

Rajarshi

unread,
Aug 15, 2007, 1:49:34 PM8/15/07
to UsefulChem

On Aug 15, 1:44 pm, jcb <jeanclaude.brad...@gmail.com> wrote:
> Duncan/Rajarshi,
> I understand that you have a simple part of this workflow working now:
> appending a closure tag to a SMILES tag. Could you please provide
> instructions here on the mailing list so my students and other
> collaborators can replicate this?
> What is the command to use on Taverna?

I started the Taverna workbench and then opened the workflow from the
File->Open menu option.

Then I clicked on the Results tab which asked me if I wanted to run
the workflow and I selected yes. A few seconds later the results are
there

Jean-Claude Bradley

unread,
Aug 15, 2007, 3:24:04 PM8/15/07
to usefu...@googlegroups.com
I'm attaching the results that I get as a pic.  I think the workflow executed without throwing an error but I don't see the actual SMILES codes.  Also how would I change the input?
--
tav1.JPG

Rajarshi

unread,
Aug 15, 2007, 3:39:42 PM8/15/07
to UsefulChem

On Aug 15, 3:24 pm, "Jean-Claude Bradley"


<jeanclaude.brad...@gmail.com> wrote:
> I'm attaching the results that I get as a pic. I think the workflow
> executed without throwing an error but I don't see the actual SMILES codes.

Strange! It works for me, see http:/rguha.ath.cx/~rguha/tavernass.png

> Also how would I change the input?

I think the workflow hard codes the SMILES file as a URL located on
the U Manchester web site

Jean-Claude Bradley

unread,
Aug 15, 2007, 3:51:30 PM8/15/07
to usefu...@googlegroups.com
Rajarshi,
I closed and tried again but ran into the same problem.
Maybe you are using a slightly different version of the workflow
getFragWithClosure2.xml
 
When I downloaded it said that it was not able to find one component but still produced the xlm file.  Could you try to download the workflow again?
 
Jean-Claude

 
On 8/15/07, Rajarshi <rajars...@gmail.com> wrote:

Rajarshi

unread,
Aug 15, 2007, 4:02:02 PM8/15/07
to UsefulChem

On Aug 15, 3:51 pm, "Jean-Claude Bradley"


<jeanclaude.brad...@gmail.com> wrote:
> Rajarshi,
> I closed and tried again but ran into the same problem.
> Maybe you are using a slightly different version of the workflow
> getFragWithClosure2.xml
>
> When I downloaded it said that it was not able to find one component but
> still produced the xlm file. Could you try to download the workflow again?

I downloaded from the link that Duncan sent out. I downloaded it again
and I didn't get any errors while downloading.

I'm running Taverna 1.5 on OS X 10.4.10

Jean-Claude Bradley

unread,
Aug 15, 2007, 4:15:31 PM8/15/07
to usefu...@googlegroups.com
I was using a different workflow - with a 2 at the end.
Following Duncan's last link on this thread:
 
I can download without error but I get a failure when I try to run (see attached)


 
On 8/15/07, Rajarshi <rajars...@gmail.com> wrote:
tav2.JPG

Jean-Claude Bradley

unread,
Aug 15, 2007, 4:41:47 PM8/15/07
to usefu...@googlegroups.com
Rajarshi,
So it turns out you were using
 
which is the original one I was using and got no SMILES codes in the results.
I redownloaded Workflow 43 and got the same error.
At this point the only difference between our systems seems to be that you are on OSX and I am on WinXP.
Mabye we can get feedback from somebody else out there with a PC?
 
Jean-Claude

 

Hiro Sheridan

unread,
Sep 24, 2007, 6:59:56 PM9/24/07
to UsefulChem
1. Can someone tell me what I'm doing wrong here?

http://www.chemspider.com/InChI.asmx?op=InChIKeyToInChI?inchi_key=BSYNRYMUTXBXSQ-WXRBYKJCCW

I got the key from here:
http://www.chemspider.com/RecordView.aspx?id=2157

I'm trying to convert an InChiKey to InChi.

2. Is it possible to go straight from InChiKey to SMILES?

Thanks,
- Andy (Hiro)

Val at ChemSpider

unread,
Sep 24, 2007, 7:28:38 PM9/24/07
to usefu...@googlegroups.com
Hi Andy

The right URL would be
http://www.chemspider.com/InChI.asmx/InChIKeyToInChI?inchi_key=BSYNRYMUT
XBXSQ-WXRBYKJCCW

You can see it at
http://www.chemspider.com/InChI.asmx?op=InChIKeyToInChI in section
describing GET method.

There is no shortcut from InChIKey to SMILES. We decided not to go into
from/to grid of operations but implement this general operation instead:
http://www.chemspider.com/OpenBabel.asmx?op=Convert

If you'd still prefer InChIKeyToSMILES operation for you application
please let us know.

Regards
Val

Tony at ChemSpider

unread,
Sep 24, 2007, 8:00:25 PM9/24/07
to UsefulChem
To add to this comment it is NOT possible to convert an InChIKey to an
InChI String. It is a one way conversion since it is a hash. InChIKeys
CAN be used to look up the associated InChI String provided that they
have been databased together...as we have done on ChemSpider.

On Sep 24, 7:28 pm, "Val at ChemSpider" <v...@chemspider.com> wrote:
> Hi Andy
>

> The right URL would behttp://www.chemspider.com/InChI.asmx/InChIKeyToInChI?inchi_key=BSYNRYMUT
> XBXSQ-WXRBYKJCCW
>
> You can see it athttp://www.chemspider.com/InChI.asmx?op=InChIKeyToInChIin section

Andrew Lang

unread,
Sep 24, 2007, 8:52:38 PM9/24/07
to usefu...@googlegroups.com
Thanks guys,
 
I am actually trying to call the webservice using php. The code below is very similar to what works for Rajarshi's web services but I can't figure out why its not working for chemspider. Of course, I added to my confusion by getting the wrong url. Here's my code (snippet):
 
$wsdl="http://www.chemspider.com/InChI.asmx?wsdl";
$client=new soapclient($wsdl, true);
$param=array('inchi_key'=>$key);
$inchi = $client->call('InChIKeyToInChI', $param);
 
All I get back is 'Array'
Does anyone know what's wrong here or have an example of some php code that consumes the web service correctly?
 
Thanks,
 
Andy

Hiro Sheridan

unread,
Sep 26, 2007, 4:27:03 PM9/26/07
to UsefulChem
I have just finished a molecule rezzer (modeled after Orac) that
accepts InChI, SMILES, and InChIKey identifiers. It takes user input
and querys the Chemspider web server (for InChiKeys) plus two more of
Rajarshi's web servers (format conversion and minimization) and then
rezzes a 3d minimized molecule in Second Life. It is limited by Second
Life memory restrictions but will produce molecules of up to about
(I'm not sure of the exact limit) 120 atoms. If anyone would like a
demonstration, just IM me in SL (Hiro Sheridan).
Andy (Hiro)

Andrew Lang

unread,
Sep 26, 2007, 9:28:16 PM9/26/07
to usefu...@googlegroups.com, ccu...@indiana.edu, ckl...@oru.edu, David Wild, Desi, dhal...@oru.edu, Eloise Pasteur, gthyv...@oru.edu, jlal...@brynmawr.edu, Kate Sellar, Kelly Bailey, Ken Hay, kneue...@selu.edu, kpre...@oru.edu, kw...@oru.edu, mlmi...@oru.edu, sc...@oru.edu, wcol...@oru.edu

My second ever you tube video. Much better than the first but still a long way from Spielberg.
 
Check it out:
http://youtube.com/watch?v=tOfhuoRbnCg&v3
 
Andy Lang
(Hiro Sheridan in SL)


> From: games...@hotmail.com
> To: usefu...@googlegroups.com
> Subject: [UsefulChem] Re: UsefulChem web services

jcb

unread,
Oct 1, 2007, 5:01:39 PM10/1/07
to UsefulChem
Hiro asked about finding enoyl reductase on ChemSpider. I am assuming
at this point that ChemSpider is only indexing organic molecule, not
biological ones. But I am posting here so CS people can comment.

Hiro, the structure of enoyl reductase can be found here:
http://www.pdb.org/pdb/explore.do?structureId=1ZW1

You can download the pdb file there. It certainly would be handy to
have your rezzer represent enzymes directly from the pdf file! But
you probably could not put every atom. If you click on the Jmol view
there you can what a ribbon structure looks like - that should not
take up too many prims.

To anybody out there - is the ribbon structure of proteins available
in a 3D coordinate file that Hiro could plug into his rezzer?

Tony at ChemSpider

unread,
Oct 1, 2007, 11:01:01 PM10/1/07
to UsefulChem
ChemSpider does not host biological molecules...it's not planned for
the near to mid-term either. There is enough to do with organic
molecules for right now. Best wishes.
Reply all
Reply to author
Forward
0 new messages