The dilemmas of success

Over the past 50 years, crystallography has transformed our understanding of crystal and molecular structure. This prospect was a driving force for early crystallographers; Dorothy Hodgkin saw crystallography as liberating chemists from the difficult and ultimately unsatisfactory task of determining molecular structure by the classical chemical approaches. The success of crystallography has made it a uniquely indispensible tool for chemistry, biology and physics. In this respect I feel incredibly lucky to be working in such a dynamic and exciting field.

Success, and the urgent demand for structural data, has brought constant innovation in both methods and technology. Small molecule structures can be solved within hours, and the time required to solve a 2000-atom protein structure is now often less than was required to solve a 20-atom structure when I was a graduate student! This very success brings problems, however. The very real skills of crystallographers can be undervalued when many (but not all) structures can be solved routinely, with little training. It also brings increased risk of errors, ranging from minor to very serious, even wrong structures.

As Greg Petsko reminded us at last year's ACA meeting, macromolecular crystallography is rapidly heading in the same direction as chemical crystallography, and many macromolecular structures will soon be solved more or less automatically. With the help of synchrotrons, MAD phasing and automated software, several have already been solved in a day or two! The emerging structural genomics programmes will accelerate this trend. What should we do to prevent our science from being seen as "routine" in future? Should we do anything?

I believe that we need to reintroduce and revitalise the teaching of crystallography at undergraduate and graduate levels (from which it has all but disappeared in many institutions). I would also advise graduate students to develop multiple skills, so that as well as being crystallographers they are also biologists, chemists and physicists, who can bring a unique structural perspective to the other things they do. We have to be vigilant for error in our journals and structural databases, and constantly develop (and apply) better validation criteria. And we have to constantly remind administrators (sometimes even colleagues) of the importance of crystallography, its interdisciplinary place in science, and its manifest success.