L. Bourhis

The anatomy of a comprehensive constrained, restrained refinement program for the modern computing environment - Olex2 dissected

An in-depth presentation is given of olex2.refine, the new refinement engine integrated in the Olex2 program.

Estimating temperature-dependent anisotropic hydrogen displacements with the invariom database and a new segmented rigid-body analysis program

A novel method and a new program for estimating anisotropic displacement parameters for H atoms are presented. Results are validated against molecular orbital computations and neutron diffraction data.

OLEX2 – a complete package for molecular crystallography

Structure Refinement can be carried out with olex2.refine. The refinement engine is based on the cctbx and provides all the functionality required for a meaningful structure refinement. A general system allows the implementation of any constraint, which has been used to provide all ShelXL constraints as well as some new constraints. All ShelX restraints have also been implemented. Of course, ShelXL can be chosen as the refinement engine as an alternative to olex2.refine.

Small-molecule refinement using the computational crystallography toolbox (cctbx) withOLEX2

improvement of the refinement can be obtained by eliminating the badly fitted reflections and/or by correcting the model, which could be too inaccurate or involve too subtle features to be modelled depending on the quality of the observations processed. A successfully recognition of the actually influencing reflections is represented by some statistical estimators based on leverage analysis (Belsey et al., 1980, Merli, 2005). Among several exploitable estimators, there are some of them that seem to be particularly suitable in crystallographic calculations: i) Cook distance (Cook, 1977), ii) DFFTIS and iii) DFBETASij (Belsey et al., 1980). In particular, estimator i) can be used to recognize an actual outlier of the refinement, whereas estimator ii) is useful to investigate the effects when a reflection is eliminated from the data set and iii) is able to recognize the variables that are mostly influenced by each reflection. The combined analysis of these diagnostics will be presented. It will be shown how successfully it yields to the recognition of a dangerous reflection and/or the inaccurate estimate of any particular variable. Belsey, Kuth & Welsh (1980) Regression diagnostics. J. Wiley & Sons Cook(1977) Technometrics, 19, 15-18 Merli (2005) Acta Cryst. A61, 471-477

Open Source software for small molecule crystallography

The ’’Next Generation of small molecule Crystallographic Software’’ is a joint project between the University of Oxford andDurhamUniversity, with an external management advisory group and scientific advisory committee of experienced senior crystallographers frommany Universities. The consortium was set up to react to the challenges facing current crystallographic software and the knowledge contained within them. To date, almost all widely used small molecule programs are written by single individuals or small groups to software standards dating from the 1970’s. The algorithm details are not documented and this software is neither extendable nor supportable. There is a risk of loss of technical knowledge as authors retire or leave the field, hence the pressing need to bridge the gap between previous and future generations of crystallographers. The aim of the project is to develop Open Source software solutions for Crystallography, which we consider to be a natural way for a project to evolve, since we adhere to software culture that encourages code-sharing. This will ensure the emergence of new science and will complement existing macromolecular crystallographic developments within the domain of small molecule crystallography. To this end, we will implement a pilot design and develop a new program to modern software standards, unhindered by legacy code, supported by detailed documentation, and which will guarantee maintenance and sustainability. It will provide a software kernel on which other researches can build new applications. The initial phase of the project will include all the fields covered by mainstream crystallographic software, such as CRYSTALS, GSAS, JANA, PLATON, SHELX, TOPAS, XP, and include data quality analysis, model building, electron density maps and related analysis, refinement, structure analyses and validation. The presentation will explain why it is important to develop such a project for crystallography now, and what the key challenges will be in completing it. The development team recognises that while current crystallographers are generally content with existing software, the next generation will requite something altogether more integrated and sophisticated, and their aim is to anticipate these needs. After introducing our conceptual view of the project, we will show how we intend to develop methods and algorithms within the new software architecture and at the same time benefit from existing programs. Other parties will be strongly encouraged to contribute their knowledge and ideas to the main development code base. As an Open Source software project, the code will be openly available to all for re-use in order to develop their own extensions. m34.p09

Small molecule crystallography for the future

Olex is a powerful molecular graphics tool. As can be expected from any molecular viewer, Olex provides fully customisable, powerful molecular graphics. Olex can interact seamlessly and intelligently with the refinement program ShelXL, setting it apart frommost other programs of its kind. The highly sophisticated refinement tools provided by ShelXL can be accessed and addressed intuitively,making building complicated structure models very easy, even for the novice user. The User Interface to Olex has a high degree of redundancy: the program can be driven from the command line only (using familiar ShelXP syntax), the context menus (right-click), as well as the highly adaptable (and customisable) Graphics User Interface. This high degree of flexibility allows the program to be accessible to all levels of user experience. m34.p01