The workshop itself was phantastic – we had 40 participants with well balanced contributions from industry and academia. The first half day was dedicated to tutorials on various aspects of CDK, basic installation, our CDK and Oracle-based chemical search engine, scripting of CDK from within Bioclipse and CDK-based workflows in KNIME. The second day brought talks about science and applications based on CDK by John van Drie, Asad Rahman und Oliver Karch.  I reported earlier on those talks.

This official part was followed by a developers workshop in the format of an unconference.  Participants suggested topics on the fly and then voted with their feet. The topics that then made it where:

• Clojure
• Mining ChEMBL with CDK pharmacophore stuff
• Threading group
• Top 10 improvements (CDK WS 2009 unconference)
• JChemPaint/Rendering

all of which reported their results on the CDK wiki.

Another important outcome was a movement to join forces in creating and maintaining joint CDK-based plugins for KNIME and Bioclipse. There is a lot of synergy to be harvested. Licensing problems will need to be overcome but that seems doable.

As anyone can read on our CDK homepage, the CDK was started because my old compchem java classes that I had written for the first version of my structure elucidator SENECA and which where also the basis for the first JChemPaint releases needed a re-write (they where actually not bad and pretty fast, but not really object-oriented). By that time, friend Egon had come on board and so we took advantage of a joint visit to Jmol creator Dan Gezelter at Universtiy of Notre Dame to start the redesign. Unfortunately, we lost the two screenshots from Dan Gezelter’s whiteboard at nd.edu where we sketched the object hierarchy which is still largely in place today.

So, here is the current status of CDK in numbers:

• In September 2009, CDK will turn 9 years old and we already start planning for the 10th anniversary workshop.
• Its code base is more than 90,000 lines of code in more than 900 classes and over 9000 methods.
• As of today, 20/04/2009, CDK on Sourceforge has 67 registered developers and
• 86 (111) people are subscribed to the cdk-devel (cdk-users) mailing lists.
• According to Ohloh, CDK is now worth $ 4.6M and took an effort of 84 person years to create.
• The CDK article published in 2003 has been cited 68 times according to Google Scholar (I could not get my VPN to work so I do not know what WebOfScience says about our current citation count)
• According to SourceForge, CDK has been downloaded 90,000 times since 2001.

On the first workshop day, we’ll have tutorials on

• How to get started with CDK, given by Egon Willighagen
• CDK and databases, by Mark Rijnbeek
• CDK and Bioclipse, by Ola Spjuth
• CDK Workflows in KNIME, by Thorsten Meinl and Bernd Wiswedel

followed by the workshop dinner and a second day with scientific talks John van Drie, Asad Rahman and Oliver Karch.

The final discussion will include a presentation by Mark Forster, Syngenta, on his observations on the usability issues of CDK in industry. We also owe Mark a lot for creating the freely distributable linux image with the CDK and all the CDK related software used at the workshop.

On Tuesday afternoon we’ll have a developers workshop, which, traditionally, has the format of an unconference. The ideas for this format are taken from a concept called “Open Space Technology“, which we only discovered after having practised them for more than 5 years . To be fair, OST conceptualizes things nicely and relieves one from figuring out how to run meetings as openly as possible. Citing from the Wikipedia articles linked above, the idea is to have a facilitated, participant-driven conference centered around a theme or purpose, in a self-organising process; participants construct the agenda and schedule during the meeting itself.

A facilitator (me in this case) introduces the principles of OST to the participants. Participants then write the title of a session they would be interested in on a piece of paper, walk to the front of the auditorium and announce the title. If a participant proposes a topic, he or she should be passionate enough about the topic to lead the respective session.

OST philosophy is based on four rules and a law:

1. Whoever comes is the right people: this alerts the participants that attendees of a session class as “right” simply because they care to attend
2. Whatever happens is the only thing that could have: this tells the attendees to pay attention to events of the moment, instead of worrying about what could possibly happen
3. Whenever it starts is the right time: clarifies the lack of any given schedule or structure and emphasises creativity and innovation
4. When it’s over, it’s over: encourages the participants not to waste time, but to move on to something else when the fruitful discussion ends

There also exists another tentative “law”, usually referred to as the “Law of Two Feet“ (or “The Law of Mobility”), which reads as follows: If at any time during our time together you find yourself in any situation where you are neither learning nor contributing, use your two feet. Go to some other place where you may learn and contribute.

Having said all this, I hope that this largest CDK workshop ever will be asuccess and most importantly fun for everyone. We’ll keep you informed at the CDK 2009 Workshop wiki page.

Image courtesy of emhuwar

The Cheminformatics and Metabolism group at the European Bioinformatics Institute (EBI) in Hinxton, Cambridge, Uk, has an opening for a phd position. The EBI is one of four outstations of the European Molecular Biology Laboratory (EMBL) and is a great place to do research in chemistry, cheminformatics and drug discovery. In all of these areas, the really exciting stuff is done at the boundaries of molecular biology, chemistry, nanoscience, etc, and EMBL is the place for such interdisciplinary research. The really cool thing and a well-hidden secret: The successful candidate will be enrolled at the University of Cambridge and eventually get a phd from UCam. Back to the records:
The Steinbeck group does research in metabolism, natural products and cheminformatics algorithm development. The successful candidate is free to choose from a range of topics (see below) or suggest his or her own project.
Applications need to be submitted by 15 July 2009 through the EMBL phd programme and ideally the candidate should clearly indicate that he/she has a preference to work with us here at the EBI.

Here are a two suggestions for potential projects

a) Structure elucidation of unknown metabolites based on mass spectrometry or proton NMR. This topic is of great importance for Metabolomics and metabolism research in general, both of which are current hot topics in molecular biology. We have a long standing history of research in this area (see our publication list) and the candidate will be able to build on existing knowledge.

b) Intelligent systems for extracting information from the chemical literature
A vast amount of knowledge is hidden in more than 100 years of chemical literature – knowledge which needs to be semantically annotated and made discoverable and interpretable by computational algorithms. Now, a scientific article in chemistry is a complex mixture of different information and data types. It contains plain text, analytical numerical data in various flavours as well as graphics of all kinds. Methods have been developed to extract of re-discover information from either of these areas. The project here aims at combining information extracted from text, tables and graphics and using each of the areas to validate data from any of the other areas.

We are glad that the EBI Industry Program as well as the EBI outreach and training department are sponsoring the event. This means that things will be much more organized as they used to be . We are going to have one day (lunch time to lunch time the next day) of tutorials on how to use CDK and talks on research using the CDK. Afterwards, we’ll have our usual free-form developers workshop for the rest of the second day with all the usual fun. Both the evening of the 20th and the 21st will be opportunities for enjoying refreshments. We also have a small nominal registration fee of 50 Pounds to cover the costs for the workshop dinner and other food and drinks.

In GIT the source code is distributed – there is no concept of a central source repository. Developers commit their chances to their local GIT systems, with all the advantages of versioning and source code history. Other developers pull code from you if they think that the changes you’ve advertised via your favourite communication channels are interesting. In theory, this allows for a very democratic and evolutionary code development. In addition to being distributed, GIT seems to be very fast when it comes to merging. Linus reports that he does hundreds of full merges per day and nothing take longer than 5 secs.

In practice, as Linus points out in his talk, there will always be one or very few repositories that people pull from – for the Linux kernel it will be Linus’s machine. In CDK it will very likely be Egon’s. Sorry Egon, you’ve got to be online all day

The last sentence already brings me to the point. I wonder if we should give GIT a try for CDK development. The advantages do sound enormous. Ok, there are disadvantage too, such as loosing the central web browsing of the SVN repository on SF. There may be ways around this, as Egon decribed here, but this seems like not using the real thing.

This is a brief impression dump after watching Linus’ talk today and I’m happy to hear your opinions

Peter Murray-Rust set the scene by emphasising the importance for Open Data. He showed some fantastic work on data extraction by OSCAR from theses, where his group had parsed a synthetic chemistry thesis into an interactive graph of a reaction network. He also showed an SVG animation of this graph as a reaction sequence, all automatically generated from an OSCAR run. Peter pointed out in the subsequent discussion that data cannot be copyrighted, which was acknowledged by all publishers in the audience. The reality is different, however, because publisher’s licenses often prevent downloading of more than few articles in a row. Detection of a robotic download for text mining comes with the danger of the whole university being disconnected. It is unclear to me how robotically parsing papers and extracting data would damage the bushiness model of publishers. It could, of course, lower the number of subscriptions from

Ian RusselL of ALPSP presented on Open Access models and how those of uses by ALPSP members. He pointed out that a lot of long-tail publishers publish only two or three journals, quite in contrast to ACS and RSC, for example. He stated that making profit is good, because it can be reinvested into innovation. I’m not sure if I’ve seen much innovation in the publishing business before the emergence of the Open Access model. He further commented on self-archiving stating that only pre-peer-review manuscripts can be self-archived without permission from the publisher. A librarian in the audience pointed out that duplication of costs by mixed read-pays and author-pays models have significantly increased the libraries expenses and Ian Russell comment was that there are no cost-savings in Open Access. Not sure if this helps. My impression is that it is not in the interest of publishers to resolve this conflict. [Editorial Note after Submission: Ian has replied to this part - See comments. ]

Robert Kiley, Wellcome Trust, summarized the Trust’s OA policy, where Trust-funded research needs to be put into pubmedcentral six month after publication. If I remember right, the Trust funds more than 90% of Biomedical research in Britain. The NIH now has a similar policy, and so has European Research Council; Robert mentioned that most text mining so far is based on PubMed abstracts, but that the full text would be required for serious efforts. He further pointed out that the number-one option for researchers to comply with the Trust’s OA policy would be to publish in a true Open Access Journal (BMC, PLOS, etc.). The second-best choice would be to publish anywhere and self-archive. The least preferable choice would be to publish with the ACS (one of the very few publishers without a Wellcome-Trust complient OA policy) and try to change the copyright notice . The Trust is in contact with publishers to make sure that authors have a wide variety of journal with open access policies to choose from. Robert highlighted the importance of OA for the long-term preservation of articles and data therein, with special emphasis on future-proofing the record of medicine. To check the compliance of authors with this new OA policy, the Trust conducted a study with 279 papers on Trust-funded research, where, if I remember right, over 90 percent of researchers were in compliance. Robert concluded his talks with mentioning UK PubMedCentral, which will be exposed for Text Mining, including Chemical Entities. The trusts next steps are to continue to work with publishers, monitor compliance of researchers, make funds available for OA and develop UkPMC

My own talk went about Chemistry at EBI and in European Bioinformatics in general.

Simon Coles, University of Southampton, talked about building repositories to preserve chemical data and publications. His view is that of an active crystallographer and he pointed out that this community could serve as a paradigm for other areas of chemistry. He pointed out that spectra are often published as inaccessible supplemental information without proper guidelines for representation. A great chance for capturing and publishing data in the very beginning are movements like OpenWetWare and Open Notebook Science which are catching and publishing laboratory experiments as they are done, and not 9 month after, when published and filtered. Simon mentioned that less than a quarter of cystral structures determined are actually published – a lot of data is just lost. This is certainly true for NMR data, an area of interest of my own research. The crystallographic community has made a step to increase the number of published crystal structures by creating Acta Cryst E (now open access) where publications just consist of the crystal structure it self and a few additional remarks. They just publish the crystal structure!!! In CIF format, computer-readable and harvestable. This would be a great step for NMR – publish structure and spectral evidence, semantically enriched, in a short communication. Simon further reported that Southampton has build the eCrystals Data Repository software based on the eGrid project, where crystallographic data can be easily deposited, supported by authoring tools, and serving as a laboratory archive. An embargo mechanism is implemented but once published, data can be harvested and analysed. The whole, funded study was really about preservation of research data, exemplified on Crystallographic data. To scale this up, Simon suggested (does is exist already?) to have a Federation of Crystallographic Data Repositories. He further mused about how that could be transferred to less well organized disciplines such as NMR, synthesis, etc. Long tail science is extremely fractioned, unorganised stuff sitting on Laptops, unorganised, inaccessible. A new kind of electronic lab notebook, the “Smart Tea Project“, can actually be taken into the lab, using mobile input devices. Analysis and discussion of on-the-fly captured data can then be published in the workgroup, between collaborators in the world, using blog-technology. Some software can enable machines and sensors to blog their results – a sensor for room temperature and or air flow in the lab for example can then be correlated with outcome of NMR experiments using time stamps.

Diana Leitch, University of Manchester, gave “the academic librarian’s perspective, in a unique way of just talking and *not* showing any slides How relieving.

I had to miss the two talks by Chris Leonhard from BMC, whose title seems to indicate growing support for Open Access and by by David Hool from the Nature Publishing Group, because I had to catch my flight from Stansted, which I also almost missed because the fantastically reliable British rail system – all trains from London to Stansted were cancelled, we were advised to take a local train to a small station in the middle of nowhere instead, where of course there was nothing like a connection to Stansted. I was one of the lucky few to get the last available taxi, driven in snail speed by an old lady Could have been very funny as a movie. Well, I made it to the gate in time, where “the machine was otherwise fully boarded and awaiting and an on-time departure” (Those of you travelling with Ryan Air regularly know what I’m referring to). I will never again be arrogant toward people showing up in Stansted sweating and on the very last minute. The all come from central London.

So, what it the bottom line from this meeting? The important message is perhaps that OA publishing has not yet quite reached chemistry but that there are grass-root movements which are going to revolutionize the way in which we publish science and scientific data, starting at the very first moment when research is performed in the lab.

Well, with the JChemPaint applet, as used in NMRShiftDB, we do have an open source structure drawing applet available but we (as in “the makers”) have to admit that it still has a pretty big foot print. If you look at Peter Ertl’s famous editor in action or the new ChemWriter editor by Rich Apodaca’s MetaMolecular, you see how fast loading improves your user experience. Our ChEBI database’s new structure search still uses a ChemAxon Marvin Sketch applet, but since we promise the world a full open source version of ChEBI, we’ll be in need for a free (as in speech) replacement.

But there is light at the end of the tunnel. Our colleague Egon Willighagen reports on his recent progress in revamping the old JChemPaint heritage into something beautiful and usable.

Peter Murray-Rust comments on the development.

Egon has built on work done by Niels Out in a Dutch Summer of Code project last year. Niels had produced both the beautiful Java 2D based renderer as well as a provisional controller.

While the details discussed there are not particularly surprising to the educated chemist, the value of the material lies in that it is a quite complete collection of cases to take into consideration. It would be a nice document to base the next JCP-rendering-engine-related summer of code on (Not sure how far Egon and Niels got last time) and there are certainly some things which could very quickly improve the quality of output from CDK SDG, the simplest example being a horizontal alignment of the molecules longest axis.

And if you think, yes, this would be a nice project for me to help those Blue Obelisk projects, please let me know