"The treatment of errors in the isomorphous replacement method"

Date:
1989-1990
Reference:
PP/CRI/I/2/24
Catalogue details

Licence: In copyright

Credit: "The treatment of errors in the isomorphous replacement method". Source: Wellcome Collection.

    1/2
    . - CC/NUMBER 49 ,-Thls week's Citation Classic IBlow D M & Crick F H C. The treatment of errors in the isomorphous replacement I method. Acte Crystallogr. 12:794-802, 1959. I [Medical Research Council Unit for Molecular Biology, Cavendish Laboratory, Cambridge, I England] A method is described for assessing the errors that arise in the isomorphous replacement method, both due to observational error and nonisomorphism. The relative probabilities of different phases may be used to calculate a best Fourier in which the errors of electron density are minimized. [The SC/® indicates that this paper has been cited in over 435 publica- Phasing X-ray Diffraction Data for Proteins D.M. Blow Blackett Laboratory Imperial College London SW7 2BZ England September 20, 1989 When I became Max Perutz's graduate student in the Cavendish Laboratory, Cambridge, in October 1954,1 was assigned a desk near the door in a large, bare room, whose walls were formed of uncovered brickwork. This room was shared by the junior mem bers of the group, often as many as eight of us. Perutz was tackling the impossible problem of finding the molecular structure of the haemoglobin molecule, using crystallographic methods. Within this problem was the simpler puzzle of assigning signs to a few hundred observed X-ray intensities to make a projection down the twofold symmetry axis of the molecule, at low resolution. Perutz 1 had al ready shown how, by introducing mercury into the crystals, most of the required signs could be found. I was convinced that this correct solution showed how to tackle the bigger problem. When a structure is viewed down a twofold axis, it looks symmetrical, and this symmetry simplifies the puzzle. In this case each X-ray intensity only needs to have its sign determined (+ or -). Any other view of the molecule is asymmetrical, and for the vast majority of X-ray intensities the phase angles can have any value between 0 and 2rr. My PhD problem, to find a general method for de termining these phase angles, was a simple geomet rical problem. 2 But there were two snags. One was that no isomorphous replacement could solve the problem on its own, because it would always give two alternative answers: the results from several dif ferent derivatives had to be combined somehow. 3 The other was that Perutz's results had shown that the effect of errors was very serious: the errors were almost as large as the differences caused by the heavy atom. One day an extrovert new arrival appeared in our room, returned from a postdoctoral year at Brooklyn Polytechnic. At first sight he was just going to be a nuisance, as his method of working was to talk loudly all the time. To me, a diffident physics student who knew next to nothing about biology, most of his con versation was gibberish. Although Francis Crick's aim was to solve the genetic code (a concept beyond me at the time), Perutz had taken him back into his lab oratory to advise on theoretical aspects of X-ray dif fraction. In due course I had to overcome my diffidence and explain to Francis what I was doing. With amazing rapidity he grasped the essence of the problem. In an hour he had shown me a new geometrical picture of the errors in phase angles and their relation to er rors in an electron density map. He had suggested a criterion for minimizing these errors and invented the name best for a map using this criterion. Following this talk, I spent months trying to gen erate explicit probability functions for the sine or co sine of a phase angle that could be integrated using this criterion to generate a best phase and weight. I failed, and instead built an analogue computer, a triangular slide rule, from which I could labori ously derive an overall phase probability curve, which could finally be used to calculate a best phase and weight. Computer programming was then in its infancy, and it was some years before I was able to make a robust computer algorithm for the method. For my PhD in 1957,1 applied this method to make one view of the haemoglobin molecule; the method has been used widely since its publication in 1959. Greater generality and rigour are quite possible, 4 but even with today's abundant computing power, the improved accuracy is small and the added com plication can be confusing. Surprisingly, the original method is still most popular. 5 Francis won a few prizes for his other work, but he got none for this. It helped me to share the Charles Leopold Meyer Prize of the French Academy of Sciences (1979) and the Wolf Prize for Chemistry (1987). replacèmenWTiethod Pmc. Roy. Soc. LondonSer. A 225:287007, 1954. (Cited 160 tim<s.) by * ^ ° S 2. Bokhoven C, Schoone J C & Bijvix-t J M. The Fourier synthesis of the crystal structure of strychnine sulphate pentahydrate. Acta Crystallogr. 4:275-80, 1951. (Cited 60 times.) isomorphous replacement. Acta Crystallogr. 9:1-9, 1956. (Cited 55 times.) 4. Sygusch J. Minimum-variance Fourier coefficients from the isomorphous replacement method by least-squares analysis. Acta Crystallogr. A-Found. Oys. 33:512-8, 1977. (Cited 10 times.) 5. Watenpaugh K D. Overview of phasing by isomorphous replacement. Meth. Enzymology 115:3-15, 1985. 16 ©1989 by ISI® CURRENT CONTENTS® THE SC VOLUME 3, NO. 21THE NEWSPAPER FOR THE SCIENCE PROFESSIONALOCTOBEK 30,1989 $3.00/£t.50 at Second Class postage ra NEWS Has Neuroscience Society Growth Been Too Fast for Its Own Good? The 14,000-Member Body Meets in Phoenix This Week! Somç Participants Fear That the Society Has Become U nwiel dy. E. Permisi EC Looks to Shut Out United-States Science - As 1992 Nears, the European Community Pushes for a Continental Focus That Could Dim United-States Hopes for Collaborati on. G.C. Anderson Bromley Tells Congress Science Must Bolster United-States Leadership - A Beefed-Up Office of Science and Technology Policy Will Help.President's New Adviser Guide Congress in Making Hard Decisions, j. Mervis Performance Biomechanics - Music to Scientists Ears Neurologists and Engineers Expect Resfeafch on Singers and Instrumentalists to Reveal Processes Behind Human Methanical Motions. J. Seiken ' UCS Marks 20 Years As Advocate for the Wise Use of Technology - MIT-Based Union of Concerned Scientists Sponsors Campus Week in New Campaign to Save the Global Environment. K. Klinger i Genome Database Booms As Journals Take the Hard Line - More Journals Are Forcing Researchers to Turn Their Data over to Genbank. An Information Revolution in the Making?. G.C. Anderson ■ > 2 Foundations Collaborate on Cognitive Neuroscience - Top Schools Compete for $1.2 Million Grants As Pew and McDonnell Launch Fresh Assault on the Mind-Body Problem. E. Pennisi 9 Glasnost Gives United-States Teens Peek at Science Journal, Kvant. . . J. Mervis 14 AutopsV'ë Fall Imperils Research - As Physicians Cease Performing This Ancient Procedure, Scientists Complain That Vital Data Are Being Lost. P. Mc Carthy ■ OPINION 16 COMMENTARY: the 1989 Lasker Award Affirms the Value of Contraceptive Research. E. Garfield■ 19 Tale of Science Rivalry Marks - Chemist's Debut As Novelist. ■Ç. Djecassi . '' RESEARCH 20 New Techniques Spur Brain Investigation. A. Grissom PROFESSION 22 Using New Science Resources: A Key to Staying Competitive. L. Reiflehrer 23 Bristol-Myers' Unrestricted Grants Fund Neuroscientists' 'Wildest' Ideas. B. Spector 25 Science Software for the Macintosh: A Sudden Abundance of Riches. D.J. Gabaldon 26 Fluorescence Photometer Measures Low Light Levels. V.R. Sheridan, W.W. Sheridan 16 Killer Bees. R.F. Chen CURRENT CONTENTS® ©1989 by ISI® PC&ES, V. 29, #49, Dec. 4, 1989 17
    No text description is available for this image
    No text description is available for this image