ANALYSIS & OPINION

Not just semantics

17 May 2007



Tim Berners-Lee's (the creator of the world wide web) dream of a web in which 'the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines' may be at the very brink of realisation, in scientific research at least. A new market report from Atrium Research devoted a whole chapter to the use of the 'semantic web' in electronic laboratory notebooks (ELNs), and report author Michael Elliot suggested that most major pharmaceutical companies are now piloting such systems.

The report, based on research at the University of Southampton in the UK, predicts a widespread acceptance within the next five years. In fact, the university's own system has been so successful it is even considering introducing it to undergraduate courses.

Jeremy Frey, a reader of chemistry at Southampton University and the author of the semantic web chapter in the report, said: 'All this is happening very rapidly. It's really starting to take off, and once people start to use it, the uptake will be much quicker.'

At its simplest, the semantic web attaches meaning and context to data, so, for example, if you were searching for JFK the president, for example, it would only return results about John F Kennedy the person, and not the airport or stadium. It adds richness and depth to the information, converting it from a skeletal data set to full-bodied knowledge.

Within science, the meaning, attached via 'tags', may link to the toxicology of samples, previous research and details of the experiments that provided the data. It is thought to be particularly useful for the biological sciences, where research frequently takes the form of long and wordy dissertations.

Many smaller laboratories are only just coming around to the idea of an electronic laboratory notebook, and the idea of putting increased burdens on their scientists to tag and link data may be discouraging. Frey, however, believes there will be automated ways to tag the data built into the data acquisition systems, and that ultimately it will reduce workload.

It is easy to see why the idea is proving to be so attractive to pharmaceutical companies. By providing natural links between related experiments, the semantic web should encourage scientists to reuse past data sets, and to communicate and collaborate on projects. In the US and many European countries the number of highly qualified scientists is rapidly decreasing, so companies are facing increasing pressure to make the most of their existing resources.

'The benefits of replacing paper notebooks are minor in comparison to the benefits of sharing knowledge,' Michael Elliot, CEO of Atrium Research told scientific-computing.com. 'It's moving the ELN away from being a collector of data to being a full knowledge-management tool. People are increasingly saying: "solve my problem – don't sell me an anagram!"'

Jeremy Frey believes that the ready access to more detailed information will also promote greater trust in data. 'One of the biggest problems at the moment is knowing where data comes from. [With the semantic web] it will be very difficult for scientists to cover their tracks,' he told scientific-computing.com. This might not always prove to be an advantage, however, raising additional legal issues for companies keen to protect their proprietary data. Laboratories would also be obliged to prove the validity of these links during patent lawsuits.

Eventually, however, the benefits should far outweigh the hassle of integrating yet another data management system, and possibly not just in the traditional hard sciences. According to Frey, even disciplines such as archaeology and the arts could be affected. 'Things are really beginning to escalate.'

The third edition of Atrium Research's report: Electronic Laboratory Notebooks: A Foundation for Scientific Knowledge Management can be found at www.atriumresearch.com

Related internet links

University of Southampton