PROJECT 5 - DUKE - STRUCTURE VALIDATION AND IMPROVEMENT FOR PROTEINS AND N. ACIDS

点击翻译按钮获取中文摘要

The overall PHENIX project seeks to create an integrated platform for rapid development of new methods in automated structure solution, resulting in a software system that can go from experimental x-ray diffraction data all the way to a minimally biased final model. Structure validation (evaluating global and local accuracy of the molecular model) is a required part of such a process. Even more importantly, if the validation criteria are available within the PHENIX system, they can help improve automated decision-making throughout the process. The Duke group developed kinemage graphics, pioneered all-atom contact analysis as a powerful new source of independent validation information, introduced new validation criteria for RNA, and significantly updated the traditional conformational criteria for proteins. This makes the MolProbity web service and related software the most complete system available for diagnosing and then actually correcting problems in macromolecular models, and these methods are well-suited to adaptation and automation within PHENIX. Project V will collaboratively pursue the following research aims: a) provide all-atom contact, geometrical, and other quality criteria within PHENIX, both at a high level for users and at a low level as functions directly accessible by the other software components; b) for the nucleic-acid tool development, contribute quality-filtered fragment libraries, RNA backbone rotamers, conformational inference from the best-determined map features, and new validation measures; c) for refinement and model completion, develop new strategies based on automation of MolProbity structure corrections, refinement of hydrogen atom contacts, and construction of rotameric-ensemble models; d) develop custom 3D and 2D kinemage graphics to display results and alternatives at chosen stages throughout the PHENIX process.