Sir

My colleagues and I were shocked to read your News report “Protein chemists favour automatic answers” (ref. 1) in which the chloride ion channel was featured prominently as an example of an important protein structure determined with the help of high-throughput techniques. In the report, Neil Isaacs of Glasgow University is quoted as saying that the chloride ion-channel structure “could not have been done without automation”.

Chloride ion channel: structure solved by traditional science.

In fact, we used no automation or high-throughput methods to solve the chloride-channel structure2. Indeed, high-throughput methods have played no part in any of the difficult ion-channel structure determinations completed in my laboratory3,4,5. Our success has rested solely on the intense focus, hard work and thoughtful approach of a small group of scientists intent on solving an important problem in biological chemistry.

I do not wish to join the debate over the wisdom of funding robotic structural biology in the United Kingdom. I do, however, wish to set the record straight concerning a misrepresentation of the science carried out in my own laboratory. The explanation for why we have made good progress is simple — we try hard to understand the proteins that we work with. We study the structure and function of ion channels in one small laboratory. I do not have a 'structure group' and a 'function group'; young scientists who come to work with me pursue the structural and functional sides of their projects as a single endeavour. If you want to solve the structure of an ion channel it helps to understand ion channels. To understand ion channels you must study their function. Our approach is not profound, it is traditional hypothesis-driven science, and it works.