FORCED COALESCENCE PHASING - A METHOD FOR AB-INITIO DETERMINATION OF CRYSTALLOGRAPHIC PHASES

Citation
Wb. Drendel et al., FORCED COALESCENCE PHASING - A METHOD FOR AB-INITIO DETERMINATION OF CRYSTALLOGRAPHIC PHASES, Proceedings of the National Academy of Sciences of the United Statesof America, 92(2), 1995, pp. 547-551
Citations number
8
Categorie Soggetti
Multidisciplinary Sciences
ISSN journal
00278424
Volume
92
Issue
2
Year of publication
1995
Pages
547 - 551
Database
ISI
SICI code
0027-8424(1995)92:2<547:FCP-AM>2.0.ZU;2-7
Abstract
A method has been developed for ab initio determination of crystallogr aphic phases. This technique, called forced coalescence phasing (FCP), is implemented on a computer and uses an automated iterative procedur e that combines real space filtering with numerically seeded Fourier t ransforms to solve the crystallographic phase problem. This approach i s fundamentally different from that of traditional direct methods of p hasing, which rely on structure invariant probabilistic phase relation ships, In FCP, the process begins with an appropriate set of atoms ran domly distributed throughout the unit cell. In subsequent cycles of th e program, these atoms undergo continual rearrangements ultimately for ming the correct molecular structure(s) consistent with the observed x -ray data. In each cycle, the molecular rearrangement is directed by a n electron density (Fourier) map calculated using specially formulated numerical seed coefficients that, along with the phase angles for the map, are derived from the arrangement of atoms in the preceding cycle . The method has been tested using actual x-ray data from three organi c compounds. For each data set, 100 separate phase determination trial s were conducted, each trial beginning with a different set of randoml y generated starting phases. Correct phase sets were successfully dete rmined in all of the trials with most trials requiring fewer than 50 c ycles of the FCP program. In addition to its effectiveness in small mo lecule phase determination, FCP offers unexplored potential in the app lication of real-space methods to ab initio phasing of proteins and ot her macromolecule structures.