Models never lie - but only if you treat them correctly
15 August 2006
There are many glamorous models in science: mathematical rather than corporeal ones, shaped to titillate those humble toilers in the vineyard who know that they will never manage to probe their intimate secrets.
James Watson, in 'The Double Helix', speaks slightingly of one of my predecessors at UCL, the geneticist Lionel Penrose. He had spent hours playing - with the help of his young son Roger - with hinged wooden cut-outs, "self-reproducing machines" with hooks and levers that slotted into each other and which could, when pushed together in the right way, link into long chains rather like DNA itself. The approach seemed to Watson childish and he returned to his (or to Rosalind Franklin's) X-rays and to the correct answer.
The Penrose playthings, though, had their day. Now we know that proteins do snap into each other to make strings - and that BSE itself emerged when one protein changed its shape and persuaded its fellows in the cell to link together, with lethal consequences.
Professor Steve Jones [UCL Biology], 'The Daily Telegraph'