PURY: a general parameter set generator for geometry of macromolecular structures
Matej Andrejašič, Vito Türk · Acta Crystallographica Section A Foundations of Crystallography · 2006
The paper of Engh and Huber [1] with description of accurate geometrical parameters of amino acid residues has set a new standard in the protein crystallography.A similar step forward in the area of nucleic acids was made by Parkinson et al. [2].As the number of macromolecular structures of complexes with ″hetero″ ligands and their variety is growing, the time has come to prepare a parameter set, which will provide parameters for the existing and coming compounds with the accuracy and precision of the Engh-Huber and Parkinson parameter sets.Since such parameter set extends over thousands of atom types and parameters connecting them, it is clear that it can only be reliably constructed, maintained and updated in an automated manner.With these goals in mind the computer program PURY was written.The program PURY operates in two modes: the crystal structure database digestion mode and the molecular parameter construction mode.In the digestion mode it generates the entire library of parameters based on high-resolution crystal structures.The program reads an unlimited number of high resolution structures (for example the whole Cambridge Crystallographic Data Base), assigns atom types taking into account the local geometry and updates the digested list.The digested list contains all bonding parameters (bond length and angles, torsional angles and improper geometries such as chirality and planarity) together with the frequency of their appearance, sum and sum of squared values.Using this description the digested list can be anytime expanded with a private structure collections, which will update, and when found necessary expand, the digested list.In the construction mode, the program reads a description of molecule and generates molecular topology entry and corresponding parameter list for refinement and energy calculations.Currently, input in the PDB form and output for MAIN, X-PLOR-CNS programs are supported.Input as well as output formats will be expanded to support interaction with a variety of computer programs used in refinement and energy calculations of macromolecules including a web server support.