Thursday, March 29, 2007

ACS Chicago - Day #3

Tuesday promised to be an interesting day: an interesting 'Scientific Communication' CINF session in the morning and early afternoon. And, rather important to me, the Blue Obelisk dinner that night, just after another CINF party, where I chatted with a few others about options of a chemistry equivalent of the Google Summer of Code; who knows what happens this summer, but start thinking about ideas on how to increase the web experience of chemistry journal web pages.

The GSoC
Now that I am talking about the GSoC, you might have realized that the CDK and Bioclipse did not make it as mentoring organization. While I had not seriously expected it, all the enthusiasm from within both projects including several interested students, I was a bit hoping for getting accepted with at least on of them. Meanwhile, KDE, as expected, is approved, and actually contains two interesting chemistry project ideas too. One is about a 3D viewer/editor for which 7 students send Google a proposal, and the other about text mining of chemical content on the desktop, using Strigi (two students). Both topics have one excellent proposal, who do good in the ranking process. So, we might have some chemistry in the GSoC afterall.

OK, back to the ACS meeting. Fahrenbach had a presentation on blogging too, but don't remember anything special about it. The CHED session was more elaborate on the whole topic, and since you are a reader of chemical blogs, you all know about this anyway :) Loney introduced which is building a social network around biotechnology. There are other community sites like this, and my major problem with these community building efforts is that they are too well defined. I much prefer to work in a more open environment where I can get in contact just as easily with people outside some specific topic. For the rest, the set of technologies is rather comprehensive.

Frenkel spoke about the imminent success of ThermoML, which is now being supported by vendors and publishers, smoothing the whole dissemination of data supported by this format. It is basically what CML is attempting to achieve in molecular structure data. Day is having a good go at this with crystal data, and his CrystalEye project is supposed to be launched next month.

Hey, at, had a rather manager level presentation, with very little value for someone into the field of 'data lifecycle and curation'. Rather disappointing on a scientific conference, or am I judging the ACS conference here? If the Microsoft is getting interested in chemoinformatics that might be a good thing, as long as the are OK with open source, open data and open standards. Who knows...

Rzepa had his presentation on the semantic wiki, which he, in similar form, held at the German Chemoinformatics Conference too. New, I think, were the sheets on reasoning based on the content of the wiki. That was rather interesting. If we all would make our chemical knowledge available as RDF, then this can become a big thing very soon. I skipped the presentation of Renear on ontologies, though it was actually one that I had hand picked; but I was simply too tired. Will watch the podcast when available. (BTW, are they making podcasts for the CINF session only, or for the whole ACS meeting?)

In the afternoon, I also followed just a subset of the presentations. The last one was by Scott who spoke earlier on Second Life. I'm really interested in seeing where this is going, though I have my reservations if this is the right medium for mining chemical knowledge. Today she spoke on the social bookmarking and podcasting initiatives at Nature and Nature Network. The latter is a social site, like like BioTechExchange, but not limited to one specific topic, and more interdisciplinary (my account at NN). I blogged about some early issues some time ago.

Jabri showed us how Jmol is adding value to the ACS Chemical Biology journal. Yes, that's what she said. An opensource tool, developed by people on their free time, is making an ACS journal more valuable. I am very happy to hear that, and it strongly supports our view that opensource chemoinformatics is very important. Some more support from established organizations might be in order indeed!

Blue Obelisk
It has become a habit to organize Blue Obelisk dinner to talk about opensource, opendata, openscience, and the future. Actually, we secretly talk about talking over the world, but I can't say that. (Neither can I tell anything about our secret rituals. The protocol for becoming member of our society is quite simple though: be clear about your opinion that ODOSOS is the future.) Dinner was great, and it was great talking to several older and newer members of the movement. Cheers all! Oh, there also was some awards involved, but I hope Peter and Christoph will blog about that and post the pictures.

ACS Chicago - Day #2

The wetter was much better today. This is a view on downtown from the walking bridge between Lake Side and McCormick buildings of the conference site:

CINF morning
Yeah, more CINF session reports; I'm a chemoinformatician, remember. Chen showed us around in the latest changes in ChemDB, such as retrosynthesis planning. Banik shows a patented method for showing differences in a set of spectra, though his examples were not really impressive; if the method is really powerful, the examples might have been picked a bit more careful. And I have to say, in retrospect, I found the presentations in the CINF sessions typically of lower quality than I had expected for the big ACS meeting. Fortunately, meeting all the people here makes more than up for that. Guha presented a potentially powerful method to cluster large and huge data sets with a method that approximated SVD by splitting up the full matrices into many smaller ones.

Laptop Lane
The idea was that us bloggers met up, but that did not quite work out. No problem though. Spoke about many things with several people. Peter pointed me to an email from Noel on getting blog comments on JACS/ChemComm/JCIM papers on the table of contents. And somewhere that day, Jean-Claude pointed me to more chemistry in Second Life (set up by Beth). Both are great follow ups on the blog/wiki session by CHED yesterday! Around 17:00 we left for one of the receptions in the W hotel in one of the WOW rooms, though the view was not that spectacular:

I did not stay very long at the party, as I had to leave for the Bulls game against Portland. That was fun indeed! If I knew my first breakfast was going to cost 22 dollar, I would have bought a better ticket, but now I had a Stand 1 ticket which is so high up in the stadium that even the security people had not idea where to send me :) The view was still more than good enough to feel the pain/joy of the blocks and dunks-in-your-face:

Monday, March 26, 2007

ACS Chicago - Day #1

I was happy to notice just a minute ago that the first blog items covering the ACS meeting are popping up: C&EN has a set up dedicated blog about the meeting, Nature's Sceptical Caterine wrote she has reached the meeting too, Richard wrote about the scent of bugs in wine (or so), and Kyle won't make it other than tomorrow. Additionally, Nature is running a coverage of the ACS meeting. On the reader side, Paul is hoping that Whitesides will be blogged about.

My first day at the conference was interesting. The huge facility makes navigation a bit problematic, and we seemed to make it a habit to explore the wrong end of the building before heading in the right direction. There are a lot of maps in the ACS On-Site Meeting Program, but a nice overview map is lacking. Anyway, I spent the morning session in the 'blog, wiki, and podcast session', and the afternoon in CINF session honoring Prof. Wiggins.

Vogel was the first speaker in the CHED C Section morning session, and spoke about blogs and RSS feeds in general. Mitch' Yahoo Pipes hackup was mentioned in one of the talks in this morning session. Currano followed with a discussion on social bookmarking, and so did Pence who focussed on the function in education. Francl put chemical blogging in some perspective which led to a short discussion on the difference in idea between blogs and wiki's. Gelder and Picione spoke about podcasting as multimedia blogs. Scott represented recent work by Nature in exploring Second Life technologies, and mentioned the chemistry on their island, which happened to host a session of the First Online EMBL PhD Symposium last year. Bradley spoke about how he integrated blogs and wiki's into there practicals. The atmosphere of the session was relaxed and the discussion lively.

The downside of all these parallel sessions is that it is bound to give clashes. It's apparently even supposed to, because the ACS website private schedule assistant is made to make you aware and resolve such clashes. So, while I had to skip the CINF morning session honoring Wiggens, I had to skip the CHED session on social networking continuing on the CHED morning session. For example, I has to miss the presentation by Rzepa on the semantic wiki (Henry, I hope to have made up for it, by plugging your work here :)

Murray-Rust was the first speaker of the CINF Section A afternoon session, and talked about mashups, text mining and other things done in Cambridge. He also mentioned recent Greasemonkey scripts using comments from and enhancing our chemical blogs, now described at the Blue Obelisk website. (Especially the Chemical blogspace enhanced TOC of chemistry journals is nice.) Wild spoke about integrating text mining and chemoinformatics tools, and showed a mockup of a 'by the way' system for PubChem, where a PubChem entry would be enhanced with 'BTW, did you know that these 7 articles mention this molecules, and that ... etc'. These things are going to happen this year. Heller held his usual talk on InChI and PubChem, though the content has slightly changed since the last two versions I've seen (not the message, though). Doman gave a practical example showing backing up earlier statements that too much information is lost in the publication process. Heritage showed Elsevier/MDL's view on the future of chemoinformatics, and accurately touched where it is currently failing. Amusingly, he pointed out that Elsevier, the publisher, would love to see more accurate QSAR/QSPAR/VS/etc models; ironically, it is, actually, for a large part caused by data not ending up in publications that predictive models are not as accurate as they could be. So, while looking at the chemoinformaticians/metricians, they should really be looking at themselves.

Some of these presentations mentioned directly or indirectly things I worked on. Thanx for doing that! Because I knew that there was funding for going to this meeting, only after the poster submission deadline was closed, I am not in the opportunity to present my work myself.

The evening is for the traditionally parties, time to eat, drink, network, make deals and try to convince others about the virtues of ODOSOS.

A last reminder: tomorrow afternoon at 13:00 at Laptop Lane in the exposition area is a meeting of chemical bloggers. Please join and chat IRL for once! :)

Sunday, March 25, 2007

Arrived in Chicago...

I arrived in Chicago yesterday afternoon. Much warmed than the cold Chicago the ACS promised me, so my winter coat was really not necessary. Is this global warming? Or was the ACS simply wrong? Anyway, very foggy indeed, just like the Chemistry World blog wrote:

There were several other Dutch chemists on the plane, among which a few formed postdocs from Nijmegen, who I knew from the time I was still a M.Sc. student in organic chemistry. The plane was nice too, a Boeing 747, the first time I flew with one. OK, there now is the new Airbus, so the Boeing has lost some of its prestige, but here's the image anyway:

The 380 is actually supposed to be in Chicago, but I did not see it. OK, going for breakfast now, and to the ACS conference site afterwards.

Thursday, March 22, 2007

Chicago (Bulls), here I come!

I had some fun today with making prints of reservations etcetera for my trip to the ACS conference in Chicago. Went over to the website to make a print of the location of the hotel I am in. (Intercontinental Chicago: in case you want to leave me a message to meet up over breakfast or so.) Anyway, so at the ACS website I found a notice that the ACS Housing people closed down and that I should contact the hotel directly. Fine, no problem. Oh wait, my hotel is not in the list. No worries, I just enter my last name and acknowledgment number. Huh, they don't know me?? Already worried about which bridge to use as backup alternative, I emailed the organization which now takes care of it, being answered some 15 minutes later that they no longer do the hotel administration for that ACS conference anymore. That indeed rang some bell; I went over back to the ACS webpage, and this time found the correct ACS housing webpage. I had been using one from a previous ACS conference. Yeah, one of my finest hours :) Things are sorted out now, as I had already email the hotel too. Things are fine, and so my nerviness activity is back to normal. (If you care to reproduce, just go to the page for International Visitors linked from the Chicago conference homepage, scroll down to "Preparing for Your ACS Meeting Experience" and click the Hotel Information link. Makes sense, because the international guests already know how things work :) And, yes, I could have seen it mention SA in the subtitle, I know.)

My ACS Schedule

My schedule is pretty regular, filled mostly with CINF and COMP presentations. The Monday and and Wednesday afternoons are empty, though there was some plans to meet up with bloggers (and here) on Monday afternoon, which I still think we should do, even though Gaussling had to back out. I hereby suggest we meet at 13:00 at Laptop Lane on the ACS Show Floor. Sunday evening there are all sorts of parties, and I and some colleagues are going to the party for international guests at the Sheraton Chicago. Monday evening is reserved for the Bulls. I am indeed some 10 years late, but happy to finally be able to visit a NBA stadium during the season. They play against Portland on Monday, while against Detroit on Tuesday. I figure that game would have been nicer, but Tuesday evening is reserved for the Blue Obelisk social event at the South Water Kitchen.

Suggestions, like these from Carmen are most welcome!

Friday, March 16, 2007

Pipelining chemical information with Yahoo Pipes

Chemists are picking up Yahoo Pipes, or, as Noel calls them, Pipeline Pilot for RSS feeds. I tend to agree, as the source of the workflows are closed, that is, at least require registering to the Yahoo webpage.

Several chemical applications have been developed since. One was developed by Kermit who wrote an aggregator for mass spectrometry journal articles. And Mitch has set up a similar feature for ACS journals.

Now, what I am really waiting for, are the first applications that deal with molecular structures, and a pipe that alerts me about publications in which molecules are discussed matching a certain MQL molecular query for an interesting substructure.

Wednesday, March 14, 2007

What is dapagliflozin?

QDIS blogged about Bristol-Myers and AstraZeneca teaming up for a new drug called dapagliflozin. Now, dapagliflozin is, this week, the most used search keyword in Google, leading to Chemical blogspace.

I wondered what the chemical structure of this compound is. The AstraZeneca and Bristol-Myers Squibb websites don't say. Since everything in pharma is patented I went to the US patent database and a search for DPP-4 AND inhibitor found the patents 6,995,183 and 6,995,180. But that does not help me either.

Does anyone know the chemical structure of this compound? Just the InChI would be fine...

Thursday, March 08, 2007

Fast molecular similarity with a new 3D shape descriptor

Jim reported about SPECTRa being in the news and ./ about Toward a 3D Search Engine. These two items have in coming that they deal with the article Ultrafast shape recognition for similarity search in molecular databases by Ballester and Richards (DOI:10.1098/rspa.2007.1823). The NewScientist wrote up their angle on it, with a quote from Henry Rzepa.

The article proposes a new shape descriptor which is requires little computational resources to be calculated. It consists of 12 numbers describing the shape, and a simple similarity measure converts it into similarities. The results shown in the article, and replicated in the NewScientist article linked above, are interesting enough for me to wonder if I could Federico, one of our CUBIC students, to work on this in the last two weeks of his practical.

BTW, Andreas, don't those review articles (viz. DOI:10.1039/b409813g) work out good for your citation count ;)