for Supplement MolProbity macromolecular model validation assigns overall quality scores, but most importantly, it reports probable local errors with graphic flags that help structural biologists fix most of those errors. It is considered state-of-the-art for model validation, has a large worldwide user base, is incorporated into most model-building and refinement software systems, and is central to model validation at the Protein DataBank. The component programs add and optimize hydrogen atoms to analyze and display all-atom contacts (H- bonds, favorable van der Waals, and serious steric clashes of unfavorable overlaps ?0.4). That distinctive feature of MolProbity is augmented with updated versions of traditional validation measures such as Ramachandran and rotamers, with RNA ribose pucker and backbone conformers, and most recently with new criteria such as CaBLAM that are especially useful at the 2.5-4 resolutions common for structures of exciting molecular machines obtained by crystallography or cryoEM. The programs in this effective and widely used toolset are diverse and rather loosely coupled, and some are more than 20 years old. For quite a while they have been in great need of a deep rewrite and modernization beyond what an academic research grant can support, in order to stay maintainable, extensible, and robust. Our problem with the MolProbity code is the sort of dilemma that this new type of supplement is designed to solve, in order to provide open scientific software that enables its users to take full advantage of the broad capabilities now possible for secure and stable accessibility, interoperability, and re-use.We believe MolProbity is worth this investment, first to the structural biology community and second to the biomedical end-users of these structures. I am very fortunate to have identified a professional contractor highly experienced in many challenging computational improvements, rewrites, and interactive 3D graphics, that is somewhat familiar with macromolecules and interested in taking on this project. The current plan is to unify our MolProbity code base around Python with calls to well-optimized and hardened libraries in C++ or C, and to use the same standard-format parsers, internal data hierarchy, in-memory communication, and open CCTBX toolbox utilities used by the most recent of our programs which are integrated with the Phenix software system. The rewrite will include more complete regression tests, develop a standardized build process, and explore a variety of cloud hosting options for the web service.
for Supplement: My lab's MolProbity validation software flags and helps fix local errors in the 3D models of macromolecules that are used to understand biology and to design drugs and vaccines. The MolProbity service works very well and is considered state-of-the-art, but its component programs have been developed over the course of 20 years and are badly in need of a thorough upgrade and unification to modern, robust, and maintainable standards. While our programmers cannot accomplish that, this supplement will provide the professional help necessary to improve and sustain this important validation service for our many worldwide users into the future.
