Protein knowledgeat a click

Molecular biologists at Birkbeck College are using the World Wide Web, Tony Durham explains. The glossary group are talking about building their own room and having meetings in there." Fortunately it will not cost much since the new room alluded to by Alan Mills, Birkbeck College crystallographer, will not be on campus but in cyberspace, that vast garden of forking paths that you enter when you log in to the Internet.

The glossary group is just one of the active, self-organising groups that have formed among the 300 students participating in the principles of protein structure course, a pioneering experiment in the delivery of knowledge through the World Wide Web.

The Web was invented by physicists, but molecular biologists have been seduced by its potential for worldwide sharing and distribution of information, both in the impenetrable codes of computer software and DNA sequences, and in vivid diagrams and images that speak to the human eye and brain. The world's store of biological data is now doubling every 18 months.

Most of the working parts of living cells are protein, and the development and improvement of drugs now depends crucially on knowledge of the three-dimensional structures of enzymes, receptors and other proteins. "This subject has enormous importance for modern biology and for molecular medicine," says Peter Murray-Rust of Glaxo Research and Development. He is a visiting professor at Birkbeck, an Internet enthusiast who has written software to help World Wide Web authors, and Dr Mills's closest collaborator in the launch of the PPS course.

The PPS course taps directly into the world's biological data banks. The Brookhaven Protein Data Bank and the European Bioinformatics Institute in Cambridge have not merely opened their data vaults to the students. Their experts are on tap to answer students' questions and contribute course material. A signpost to the PPS course's "hypertree" of Web pages appears on Brookhaven's own World Wide Web home page. The Brookhaven data bank may not be doing this for purely altruistic motives. Its grant is up for renewal, and visible evidence of its contribution to the biological community could strengthen its case.

Students on the course range from undergraduates, who may be unable to find an equivalent course at their own university, to scientists from pharmaceutical companies seeking a mid-career boost. "This is a very important part of continuing education for the biotechnology companies," Dr Mills points out. Birkbeck was founded specifically to provide adult education.

The experiment is sponsored by the Globewide Network Academy, a non-profit organisation with headquarters in Austin, Texas, which is fostering experiments in higher education on the Internet.

Distance learning is a lonely business, only partially alleviated by being able to see pictures of other people, "they can just click on my name and see what I look like", says Dr Mills, conjuring up his portrait with a flourish of the mouse.

There are also communal electronic meeting places such as BioMOO. BioMOO is a software creation from the Weizmann Institute in Israel, which has been adopted as the course's virtual student union building. You go there (you hit keys and click mouse buttons), you meet people (the writing on your screen tells you who is present), you chat with them (more writing on screens) and if you have mastered basic BioMOO skills you can think about modifying BioMOO itself: building a new conference room, for example. Going, meeting, chatting and building are all figures of speech for activities conducted from chair and PC. The Internet is no place for people with low metaphor tolerance.

The course uses what Dr Mills believes to be a unique combination of tools to extend the basic publishing and browsing capabilities of the Web. Molecular structures are encoded in the new Chemical MIME formats compatible with most of the molecular modelling and 3D viewing software used by chemists. Students register for the course by filling in forms on their screens, a recently-added WWW feature which is also much used by businesses operating on the Internet. Most of the course material is publicly available to all Internet users, but registration brings extra privileges such as participation in discussions through the course's electronic mailing list.

There are clickable maps: use the mouse to point and click on an image of a protein molecule, and you can summon up information on a specific part of the structure. And anyone on the course can contribute definitions to the "collaborative hyperglossary" of protein-structure terminology.

The glossary solves a problem which hits most Web authors sooner or later. Technical words and phrases such as "alpha helix" and "disulphide bridge" are likely to occur repeatedly in any writing about protein structure. It is helpful to the reader if they are highlighted and linked, through the Web, to definitions which can be brought up on the screen. But it is laborious for the author to have to link each occurrence of a phrase to its definition by hand.

Just as word-processors now have automatic indexing, the PPS hyperglossary software will automatically highlight technical terms in a Web document, and link them to their definitions. You send your HTML file (the standard format for Web pages) to the glossary, which lives on a central server, and the file comes back studded with references to glossary definitions. The on-screen appearance of the document is unaltered, apart from the highlighting which shows that certain words are linked to the glossary or to other Web pages.

The economics of the experimental course are intriguing. "This has a significant cost at Birkbeck but nowhere else," Professor Murray-Rust explains. "The rest is people's time." Birkbeck is funding Dr Mills for a year to work full-time on the experiment. Otherwise, the two heaviest costs are administration and technology support.

The course has benefited from free software, and free use of spare capacity on university and research institute computers and the Internet. Intellectually, the course is equally self-levitating. Dr Mills provided a kernel of course material but the rest has been provided freely by the 300 students and a number of expert consultants. The students entered into a learning contract which obliges them to give as well as take. In practice most of the good things in the PPS hypertree - the insight-giving graphics, the lucid explanations, the skilful definitions of difficult terms - result from the work of an active core of about 5 per cent of course participants. There is no formal review process, but poor-quality material would swiftly attract electronic criticism.

What is not clear is whether this model could be extended to complete degree courses, and how it could evolve into being economically self-sustaining rather than piggy backing on existing campus-based activities and resources. It would be tempting to use the technology ruthlessly to boost student numbers and reduce faculty workload, but Professor Murray-Rust is emphatic: "This is not a way of producing low quality mass education at near zero cost."

The PPS course has attracted online students from as far afield as China and Colombia. David Moss, chairman of Birkbeck's crystallography department, is keen to develop distance learning, so that the college can serve students who live beyond commuter range. But he adds, "We would still, I think, want to see our students here and know them personally." Birkbeck's constituency might realistically expand from the London area to cover much of northwest Europe.

Students on the PPS course will receive a certificate of participation but no formal diploma. One problem is that it is difficult to be absolutely sure of the identity of a student whose only verifiable attribute is an email address. But next year some of the material will be used in Birkbeck's undergraduate courses, and eventually Dr Moss expects to provide degree courses at a distance.

The principles of protein structure home page is at http://www.cry st.bbk.ac.uk/PPS/index.html