ISSN:
1600-5724
Source:
Crystallography Journals Online : IUCR Backfile Archive 1948-2001
Topics:
Chemistry and Pharmacology
,
Geosciences
,
Physics
Notes:
The phase problem can be cast into a form equivalent to that of certain optimization problems arising in operations research. Computational techniques of mathematical programming developed for these problems are thereby made applicable to phase determination. This reformulation of the phase problem is derivable by combining the nonnegativity property of electron-density distributions with molecular model building. If a structural model is available, the algorithms select phases for the nonnegative electron-density distribution, consistent with the measured intensities, that best matches the model. If no model is introduced, the algorithms produce some (or all) of the nonnegative distributions consistent with the data. The approach to the phase-problem presented here is similar to that of R.J. Dakin (Acta Cryst. 1970, B26, 2112). However, practical computational tests also are reported: an example, the 0kl projection of dihydrouracil demonstrates that alorithms based on mathematical programming techniques can solve a real structure using real data. The coupling of phase determination with molecular model building should be particularly useful in crystallographic studies of biological macromolecules. The algorithms are well-adapted to the introduction of low-resolution models inferred from hydrodynamic measurements, or detailed structures suggested by model building or computed by energy minimization.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1107/S0567739472000129
Permalink
