Example input would be: This is a list of ethanol, methanol, and 2,4,6-trinitrotoluene.

The plain text input can be linked to the pdf2text SADI service, and the CML is suitable for the CDK-Taverna plugin, which is currently being updated by Andreas, Achim, and Christoph for Taverna 2.2. As soon as the update site is properly working, I will upload a demo workflow to MyExperiment.org.

I guess the first next activity (node in the workflow) will be around the dictionaries, as the OPSIN activity converts only IUPAC names into connection tables. I was told OPSIN parses 97% of the IUPAC names it finds, and when it does, it does almost 100% correct. Want to challenge the code? Use this web service.

He asked me to test it, and I installed a fresh Taverna install and the new plugin. After that, I used the MyExperiment plugin to download one of the CDK-Taverna workflows Thomas has on MyExperiment, and tuned it a bit to use some local input instead of the database. I took some screenshots while at it, and will use those now to talk you through the installation of Taverna and the CDK-Taverna plugin.

Download Taverna

Taverna 1.7.2 can be downloaded from this download page, but I took the Linux version from the SourceForge download site. I cannot detail the OS/X or Windows installation, but on Linux you simply unzip the downloaded file, and you’re ready to go:

$ cd taverna-1.7.2/
$ sh runme.sh

Plugin Installation

Plugins can be installed using with the Plugin manager which can be accessed via the Tools menu:

Clicking the Find New Plugins takes you to a second dialog listing known plugin sites, and the default download has several already:

The CDK-Taverna update site is available at http://cdk-taverna.de/plugin/, and we can make Taverna aware of this update site by clicking the Add Plugin Site button:

After filling out these values and approving it with the OK button, it will show up on the dialog showing all available plugins, where you need the check the check box in front of the CDK-Taverna plugin name, as done in this screenshot:

You can then hit the Install button after which the plugin will be downloaded:

After it is done downloading the plugin, you can close the Plugin Sites and Plugin Manager dialogs. I shutdown and restarted Taverna with sh runme.se, but not entirely sure this is needed. After that, the CDK nodes showed up in the list of Taverna processors:

MyExperiment Plugin

Using the same Taverna Plugin Manager you can also install the MyExperiment plugin that allows you to search, browse, preview and download Taverna workflows from the MyExperiment website from within Taverna itself. I installed the plugin, and then used it to search for CDK workflows (and downloaded a QSAR workflow):

This about everything to get you going. It’s not particularly rocket science, but I guess this howto is useful as you get to see what you should expect when setting up a CDK-Taverna environment. If you have further questions, please leave those in the comments section, and I’ll try to merge in answers where possible, or otherwise in the reactions too.

I promised details, but realize I have actually already posted a lot of them in October :

Johannes ideas led to the IO-DATA proposal (XEP-0244), which is currently marked experimental and being discussed on the ws-xmpp mailing list. He gathered a few people around him to get it going, resulting in working stuff! Yeah!

Joerg asked Could you post more results, what is it, why do we need it, e.g. why are you mentioning SOAP and cloud? Do not know enough to see the bonus right now.

What is it? IO-DATA is a protocol on top of the XMPP protocol to allow machine-to-machine communication. Actually, much like SOAP, RPC, and other platforms. How IO-DATA differs lies to some extend to the transport layer: instead of using HTTP, it used the XMPP transport protocol, also used for Jabber chat clients. It basically allows clients like Taverna to chat with services running elsewhere.

Why do we need it? Most services run over HTTP, making them web services. This is convenient, because there is much infrastructure around, like web browsers. REST services also take advantage of this. However, for heavy computing this sometimes leads to problems. For example, routers are known to have time outs on HTTP connections. To solve this, SOAP services often introduce a polling mechanism. IO-DATA takes a different approach. Instead of having to ask all the time how a calculation is doing, you can just wait for the service to send you a message when it is done. Instead of working around the lack of asynchronous aspects, IO-DATA introduces these in the protocol.

Other interesting features include that the IO-DATA integrates the interface formats for services into the service itself, SOAP needs WSDL for this, and that it features service discovery via DISCO. The latter is done with SOAP too, for example with UDDI and BioMoby. The latter also adds strong data typing for input and output of services.

IO-DATA addresses the data typing by allowing asking the service what XML Schema it uses for input and output. While XML Schema has alternative, and which may be prefered in some situations, it does allow strong data typing and supports a lot of formats in life sciences (which I’ll summarise soon).

Moreover, if there just happens not to be a suitable schemata around, you can just define one yourself, which can be as simple as a single element wrapper around some custom text-based format. You worry about supporting many formats? Well, no need. Johannes’ xws4j library, which I used for the Taverna plugin too, allows compiling a Java binding code. Bioclipse’s script environment allows you do to this on the fly: you find a service, ask for the schema, compile bindings for input and output, set up the input with the input binding, send it of to the service, and use the output binding for convenient access to the computation results. Without having to reboot Bioclipse. Isn’t that cool? Can your software do that? (See this example Gist: the io factory creates the binding).

Why do I mention SOAP and the cloud? It should be clear from the above why I mention SOAP: it offer the same functionality, but more conveniently, we think. I mention cloud here, to refer to cloud computing which is doing computation on the cloud, which is a synonym for the internet (see Cloud Computing @ Wikipedia). Because it does not use HTTP, we do not feel we can call it web service. Instead, cloud computing is a more general term, not tied to any particular architecture. IO-DATA is just one possible architecture, one we think is promising for life science applications.

Details will follow, but here’s the source code.

I uploaded the source to MyExperiment, so anyway can play with it. There is much to improve, such as using CDK-Taverna for further analysis of the results.

I am not sure if opening the workflow in your Taverna installation will automatically set up the WDSL scavenger for the ChemSpider services, which are available in a HTTP version too, btw. If not, right click on the Available Processors folder, and pick Add new WDSL scavenger… and point it to the URL http://www.chemspider.com/MassSpecAPI.asmx?WSDL. The result should look like:

Oh, and please note this comment:

These services are offered free of charge to our users during this period of testing, validation and feedback. Some of these services will be made available commercially in the future and we are proactively informing you of our intention to do this. It is likely that these services will remain available to academia at no charge. Please contact us at feedbackATchemspiderDOTcom with feedback and questions.

So, I do not know when my workflow will stop working.

After lunch Mark Wilkinson introduced BioMoby, which we actually use in Wageningen already. I have tried to use jMoby to set up services based on the CDK, but failed sofar. Will talk with Mark on that. Next was my presentation, and I spoke about CDK-Taverna, Bioclipse and some peculiarities with chemoinformatics workflow, like the importance with intermediate interaction, the need to visualize the data and complex, information rich data. Bioclipse is seeing an integration of BioMoby and of Taverna.

After the coffee brake Marco Roos spoke about myExperiment and his work on text mining. I unfortunately missed this presentation, as I was meeting with people from the EBI who work on the MACiE database (see this blog item ).

A discussion session afterwards introduced a few more Taverna uses, and encountered technical problems. Taverna2 is actually going to be quite interesting, with a data caching system between work processors, and a powerful scheme of annotation of processors, which will allow rating, finding local services, etc. More on that tomorrow. Dinner time now :)

Paul Fisher just presented his PhD work on using workflows to improve the throughput of QTL matching against pathway information and phenotype. One interesting note was its function to make biological informational studies more reproducible. He had getting the versions of online databases explicitly in the workflow, so that it gets stored in workflow output.

The Monday/Tuesday after that, I will present CDK-Taverna integration I worked on in 2005 (see e.g. Taverna on Classpath and CDK-Taverna fully recognized ) at the Taverna meeting, before Thomas continued on that, leading to the cdk-taverna.de plugin website. If time permits, I will prepare an example workflow from metabolomics. Unlike previous times I went to Cambridgeshire, I won’t fly in on Stansted, but take the EuroStar instead. I am very much looking forward to that. Unfortunately, I will not have time to visit Cambridge itself, this time :(

Today, Miguel Rojas and Stefan Kuhn discussed their research. Miguel showed the state of mass spectrum prediction using the CDK and the MEDEA plugin for Bioclipse. Stefan demonstrated the NMRShiftDB and a new lab systems for NMR experiment scheduling and management system based on that. Dr. Ott (Nijmegen, Netherlands) showed the BioMeta Database which contains metabolite and reaction information derived from the KEGG, but which fixes a set of chemical problems in the latter (see also the article, DOI:10.1186/1471-2105-7-517).

The afternoons of CDK workshops traditionally have discussion sessions and hackathons. Two groups were formed: one consisted of the KNIME guys who, together with Miguel and Federico focused in QSAR descriptor calculations in KNIME, while Stefan, Martin and me looked at the fingerprinter peculiarities that Martin found (see also this CDK News article), and came up with a possible further performance improvement of the AllRingsFinder. Because one class of molecules that is causing trouble consist of two ring systems connected by a long linker, like Choloyl-CoA (below), we anticipate that splitting the molecule up into ring systems prior to using the SSSR algorithm should speed up the complete all-ring finding process.

Currently, the spanning tree is calculated before deciding on using the SSSR finder, which, we think, can be used to partition the molecule into separate ring systems. On each of them, then, the further steps of the ring search can be applied.

After dinner (pasta/pizza), during the Spanish-German handball game, we continued the hacking and discussions, now focusing as a whole group on QSAR descriptors in KNIME. We looked at each descriptor and decided if it should go into a QSAR calculator node, or even in a node of its own.

Bugs found

I won’t close this blog entry without giving a list of problems we found in the current CDK; some minor and small, some more troublesome. Here goes: typos all over the place; the OrderQueryBond lack a return statement in an else clause; the Mol2Reader does not mark atom and bond aromaticity properly and reads a single bond as aromatic, and an aromatic bond as single; the Renderer2D does not always highlight both atoms when hovering over a bond; SmilesGenerator.parseBond() should output bond orders correctly; the SSSR finder seems to have a messed up if-else statement for the ringBondCount limit of 37; the BondCount descriptor should count all bonds by default, not just the single bonds; IDescriptor.getParameters() should return null instead of Object[0]; several programs use the SYBYL atomtype S.o2, while the specification and the CDK config defines S.O2; the IP descriptor now returns a variable length descriptor.

The conference is in honor of the life work by Prof. Gasteiger, who gave an overview of chemoinformatics in his group, Germany and Europe. He stressed the need of education in chemoinformatics, like in Obernai. He also highlighted that we, today, are still solving the same problem as 30 years ago. Which is true, which is why this channel is called Chem-bla-ics , trying to solve that problem. When asked if opensource chemoinformatics form the start would have addressed this, he replied that he requires people to cooperatively do research with his group; opensource clearly cannot enforce that.

Day 2

Todays program had a number of interesting presentations (I, unfortunately, missed the first presentation, so have to visit that group soon now, to make up for that.) Prof. Aires-de-Sousa showed his work on MOLMAP for mapping metabolic networks (KEGG really, see my earlier blog ), and showed, just as proof of principle, classification of organisms based on this.

J. Weisser talked about docking, still an obligatory topic. This work really showed two new approaches: the use of QM partial charges (the example showed an improvement in RMSD of a factor 10, not very statistical, but promising indeed); the second was the fact that water does not like to be in tight spots, because of reduced possibilities for hydrogen bonding. A concept common in understand supramolecular phenomenon, but I have not seen this applied to docking before. But I am no expert in that field. M. Wagner showed work on using KEGG data to estimate likely metabolites, and the use in reducing effects of metabolic degradation. T. Schroeter introduced me to gaussian processes, a new data modeling method. Quite embarrassing to get introduced to such, as being specialized in modeling methods for chemical problems.

The poster session was, as normally, really exhausting, talking to a lot of people. Having a booth at the exhibition on opensource chemoinformatics added a nice twist to this. I therefore skipped the FIZ-award winner lectures, so I hope someone else will blog about those.

One last note: Sun started releasing their Java platform under the GPL license. Jim, seems that they proved me wrong . The class library is still not GPL, but is expected to become licensed such somewhere in the first half of next year.