The growth of linked up data in chemistry – and good community projects

It’s been an interesting week or so in the Chemistry online world. Following on from my musings about data services and the preparation I was doing for a talk the week before last I asked Tony Williams whether it was possible to embed spectra from ChemSpider on a generic web page in the same way that you would embed a YouTube video, Flickr picture, or Slideshare presentation. The idea is that if there are services out on the cloud that make it easier to put some rich material in your own online presence by hosting it somewhere that understands about your data type, then we have a chance of pulling all of these disparate online presences together.

Tony went on to release two features, one that enables you to embed a molecule, which Jean-Claude has demonstrated over on the ONS Challenge Wiki. Essentially by cutting and pasting a little bit of text from ChemSpider into Wikispaces you get a nicely drawn image of the molecule, and the machinery is in place to enable good machine readability of the displayed page (by embedding chemical identifiers within the code) as well as enabling the aggregation of web based information about the molecule back at Chemspider.

The second feature was the one I had asked about, the embedding of spectra. Again this is really useful because it means that as an experimentalist you can host spectra on a service that gets what they are, but you can also incorporate them in a nice way back into your lab book, online report, or whatever it is you are doing. This has already enabled Andy Lang and Jean-Claude to build a very cool game, initially in Second Life but now also on the web. Using the spectral and chemical information from Chemspider the player is presented with the spectrum and three molecules; if they select the correct molecule they get some points, if they get it wrong they lose some. As Tony has pointed out, this is also a way of crowdsourcing the curation process – if the majority of people disagree with the “correct” assignment then maybe the spectrum needs a second look. Chemistry Captchas anyone?

The other even this week has been the efforts by Mitch over at the Chemistry Blog to set up an online resource for named reactions by crowdsourcing contributions and ultimately turning it into a book. Mitch deserves plaudits for this because he’s gone on and done something rather than just talked about it and we need more people like that. Some of us have criticised the details (also see comments at the original post) of how he is going about it but from my perspective this is definitely criticism motivated by the hope that he will succeed and that by making some changes early on, there is the chance to get much more out of the contributions that he gets.

In particular Egon asked whether it would be better to use Wikipedia as the platform for aggregating the named reaction; a point which I agree with. The problem that people see with Wikipedia is largely that of image. People are concerned about inaccurate editing, about the sometimes combative approach of senior editors that are not necessarily expert in the are. Part of the answer is to just get in there and do it – particularly in chemistry there are a range of people working hard to try and get stuff cleaned up. Lots of work has gone into the Chemical boxes and named reactions would be an obvious thing to move on to. Nonetheless it may not work for some people and to a certain extent as long as the material that is generated can be aggregated back to Wikipedia I’m not really fussed.

The bigger concern for us “chemoinformatics jocks” (I can’t help but feel that categorising me as a foo-informatics anything is a little off beam but never mind (-;) was the example pages Mitch put up where there was very little linking back of data to other resources. So there was no way, for instance, to know that this page was even about a specific class of chemicals. The schemes were shown as plain images, making it very hard for any information aggregation service to do anything useful. Essentially the pages didn’t make full use of the power of the web to connect information.

Mitch in turn has taken the criticism offered in a positive fashion and has thrown down the gauntlet; effectively asking the question, “well if you want this marked up, where are the tools to make it easy, and the instructions in plain English to show how to do it?”. He also asks, if named reactions aren’t the best place to start, then what would be a good collaborative project. Fair questions, and I would hope the Chemspider services start to point in the right direction. Instead of drawing an image of a molecule and pasting it on a web page, use the service and connect to the molecule itself, this connects the data up, and gives you a nice picture for free. It’s not perfect. The ideal situation would be a little chemical drawing palette. You draw the molecule you want, it goes to the data service of choice, finds the correct molecule (meaning the user doesn’t need to know what SMILES, InChis or whatever are), and then brings back whatever you want; image, data, vendors, price. This would be a really powerful demonstration of the power of linked data and it probably could be pulled together from existing services.

But what about the second question? What would be a good project? Well this is where that second Chemspider functionality comes in. What about flooding Chemspider with high quality, electronic copies, of NMR spectra. Not the molecules that your supervisor will kill you for releasing, but all those molecules you’ve published, all those hiding in undergraduate chemistry practicals. Grab the electronic files, lets find a way of converting them all to JCamp online, and get ’em up on Chemspider as Open Data.

Defining error rates in the Illumina sequence: A useful and feasible open project?

Panorama image of the EBI (left) and Sulston Laboratories (right) of the Sanger Institute on the Genome campus in Cambridgeshire, England.

Regular readers will know I am a great believer in the potential of Web2.0 tools to enable rapid aggregation of loose networks of collaborators to solve a particular problem and the possibilities of using this approach to do science better, faster, and more efficiently. The reason why we haven’t had great successes on this thus far is fundamentally down to the size of the network we have in place and the bias in the expertise of that network towards specific areas. There is a strong bioinformatics/IT bias in the people interested in these tools and this plays out in a number of fields from the people on Friendfeed, to the relative frequency of commenting on PLoS Computational Biology versus PLoS ONE.

Putting these two together one obvious solution is to find a problem that is well suited to the people who are around, may be of interest to them, and is also quite useful to solve. I think I may have found such a problem.

The Illumina next generation sequencing platform developed originally by Solexa is the latest kid on the block as far as the systems that have reached the market. I spent a good part of today talking about how the analysis pipeline for this system could be improved. But one thing that came out as an issue is that no-one seems to have published  detailed analysis of the types of errors that are generated experimentally by this system. Illumina probably have done this analysis in some form but have better things to do than write it up.

The Solexa system is based on sequencing by synthesis. A population of DNA molecules, all amplified from the same single molecule, is immobilised on a surface. A new strand of DNA is added, one base at a time. In the Solexa system each base has a different fluorescent marker on it plus a blocking reagent. After the base is added, and the colour read, the blocker is removed and the next base can be added. More details can be found on the genographia wiki. There are two major sources of error here. Firstly, for a proportion of each sample, the base is not added successfully. This means in the next round, that part of the sample may generate a readout for the previous base. Secondly the blocker may fail, leading to the addition of two bases, causing a similar problem but in reverse. As the cycles proceed the ends of each DNA strand in the sample get increasingly out of phase making it harder and harder to tell which is the correct signal.

These error rates are probably dependent both on the identity of the base being added and the identity of the previous base. It may also be related to the number of cycles that have been carried out. There is also the possibility that the sample DNA has errors in it due to the amplification process though these are likely to be close to insignificant. However there is no data on these error rates available. Simple you might think to get some of the raw data and do the analysis – fit the sequence of raw intensity data to a model where the parameters are error rates for each base.

Well we know that the availability of data makes re-processing possible and we further believe in the power of the social network. And I know that a lot of you guys are good at this kind of analysis, and might be interested in having a play with some of the raw data. It could also be a good paper – Nature Biotech/Nature Methods perhaps and I am prepared to bet it would get an interesting editorial writeup on the process as well. I don’t really have the skills to do the work but if others out there are interested then I am happy to coordinate. This could all be done, in the wild, out in the open and I think that would be a brilliant demonstration of the possibilities.

Oh, the data? We’ve got access to the raw and corrected spot intensities and the base calls from a single ‘tile’ of the phiX174 control lane for a run from the 1000 Genomes Project which can be found at http://sgenomics.org/phix174.tar.gz courtesy of Nava Whiteford from the Sanger Centre. If you’re interested in the final product you can see some of the final read data being produced here.

What I had in mind was taking the called sequence, align onto phiX174 so we know the ‘true’ sequence. Then use that sequence plus a model with error rates to parameterise those error rates. Perhaps there is a better way to approach the problem? There are a series of relatively simple error models that could be tried and if the error rates can be defined then it will enable a really significant increase in both the quality and quantity of data that can be determined by these machines. I figure splitting the job up into a few small groups working on different models, putting the whole thing up on google code with a wiki there to coordinate and capture other issues as we go forward. Anybody up for it (and got the time)?

Related articles

More on the science exchance – or building and capitalising a data commons

Image from Wikipedia via ZemantaBanknotes from all around the World donated by visitors to the British Museum, London

Following on from the discussion a few weeks back kicked off by Shirley at One Big Lab and continued here I’ve been thinking about how to actually turn what was a throwaway comment into reality:

What is being generated here is new science, and science isn’t paid for per se. The resources that generate science are supported by governments, charities, and industry but the actual production of science is not supported. The truly radical approach to this would be to turn the system on its head. Don’t fund the universities to do science, fund the journals to buy science; then the system would reward increased efficiency.

There is a problem at the core of this. For someone to pay for access to the results, there has to be a monetary benefit to them. This may be through increased efficiency of their research funding but that’s a rather vague benefit. For a serious charitable or commercial funder there has to be the potential to either make money, or at least see that the enterprise could become self sufficient. But surely this means monetizing the data somehow? Which would require restrictive licences, which is not at the end what we’re about.

The other story of the week has been the, in the end very useful, kerfuffle caused by ChemSpider moving to a CC-BY-SA licence, and the confusion that has been revealed regarding data, licencing, and the public domain. John Wilbanks, whose comments on the ChemSpider licence, sparked the discussion has written two posts [1, 2] which I found illuminating and have made things much clearer for me. His point is that data naturally belongs in the public domain and that the public domain and the freedom of the data itself needs to be protected from erosion, both legal, and conceptual that could be caused by our obsession with licences. What does this mean for making an effective data commons, and the Science Exchange that could arise from it, financially viable? Continue reading “More on the science exchance – or building and capitalising a data commons”

The science exchange

How do we actually create the service that will deliver on the promise of the internet to enable collaborations to form as and where needed, to increase the speed at which we do science by enabling us to make the right contacts at the right times, and critically; how do we create the critical mass needed to actually make it happen? In another example of blog based morphic resonance there has been a discussion a discussion over at Nature Networks on how to enable collaboration occurred almost at the same time as Pawel Szczeny was blogging on freelance science. I then hooked up with Pawel to solve a problem in my research; as far as we know the first example of a scientific collaboration that started on Friendfeed. And Shirley Wu has now wrapped all of this up in a blog post about how a service to enable collaborations to be identified might actually work which has provoked a further discussion. Continue reading “The science exchange”

Bursty science depends on openness

An example of a social network diagram.Image via Wikipedia

There have been a number of interesting discussions going on in the blogosphere recently about radically different ways of practising science. Pawel Szczesny has blogged about his plans for freelancing science as a way of moving out of the rigid career structure that drives conventional academic science. Deepak Singh has blogged a number of times about ‘bursty science‘, the idea that projects can be rapidly executed by distributing them amongst a number of people, each with the capacity to undertake a small part of the project. Continue reading “Bursty science depends on openness”