ABSTRACT
A Monte Carlo procedure is described that generates random structure factors with simulated errors corresponding to an X-ray data set of a protein of a specific size and given heavy-atom content. The simulated data set can be used to estimate Bijvoet ratios and figures of merit as obtained from SAD phasing routines and can be used to gauge the feasibility of solving a structure via the SAD method. In addition to being able to estimate results from phasing, the simulation allows the estimation of the correlation coefficient between |DeltaF|, the absolute Bijvoet amplitude difference, and FA, the structure-factor amplitude of the heavy-atom model. As this quantity is used in various substructure-solution routines, the estimate provides a rough estimate of the ease of substructure solution. Furthermore, the Monte Carlo procedure provides an easy way of estimating the number of significant Bijvoet intensity differences, denoted as the measurability, and is proposed as an intuitive measure of the quality of anomalous data.
Subject(s)
Crystallography, X-Ray/methods , Proteins/chemistry , Computer Simulation , Fatty Acid Synthases/chemistry , Monte Carlo Method , Thiolester Hydrolases/chemistryABSTRACT
Methods for automated identification and building of protein-bound ligands in electron-density maps are described. An error model of the geometrical features of the molecular structure of a ligand based on a lattice distribution of positional parameters is obtained via simulation and is used for the construction of an approximate likelihood scoring function. This scoring function combined with a graph-based search technique provides a flexible model-building scheme and its application shows promising initial results. Several ligands with sizes ranging from 9 to 44 non-H atoms have been identified in various X-ray structures and built in an automatic way using a minimal amount of prior stereochemical knowledge.
Subject(s)
Crystallography, X-Ray , Ligands , Models, Molecular , Proteins/chemistry , Adenosine Monophosphate/chemistry , Adenosine Triphosphate/chemistry , Algorithms , Benzamidines/chemistry , Cholic Acid/chemistry , Dihydroxyacetone Phosphate/chemistry , Electrons , Hydrogen Bonding , NAD/chemistry , Progesterone/chemistry , Protein Conformation , Tretinoin/chemistryABSTRACT
The relation between a Gaussian perturbation of the atomic positional parameters and the average squared structure-factor amplitude is presented. Using an error-dependent radial distance distribution of an atomic protein model, it can be shown that the Debye effects diminish exponentially as a function of increasing positional errors. These relations can be used to estimate the quality of an atomic model and the corresponding phases. The limiting case of equal atoms with an infinitely large coordinate error results in the classical Wilson model.
Subject(s)
Models, Statistical , Statistics as Topic/methods , Amino Acid Sequence , Crystallography, X-Ray , Molecular Conformation , Molecular Structure , Monte Carlo Method , Normal DistributionABSTRACT
The analytical expression for the distribution of an interatomic distance resulting from a known error-free distance and a Gaussian perturbation of the atomic coordinates is presented. This is used to estimate the coordinate error on the basis of known geometric features of protein models via the nearest-neighbour or the radial distance distribution. A simple relation is presented that describes the dependence of the map correlation on the positional error of the protein model, the resolution of the X-ray data and the overall atomic displacement parameter. The distribution of geometrical features and the relation between the map correlation and the positional error can be used in assisting the decision-making process during automated model-building procedures.