| Summary: | This thesis is concerned with the development and application of AMPLE, a software pipeline for macromolecular crystallographic Molecular Replacement, to different classes of proteins. The ability of AMPLE to solve protein structures via Molecular Replacement was first explored with two new classes of proteins: coiled-coils and transmembrane helical proteins. The results were very positive, with AMPLE solving 75 of 94 (80%) of the coiled-coil and 10 of 15 (67%) of the transmembrane protein test cases. In both cases the performance of AMPLE was benchmarked against a library of ideal helices. The performance of idea helices was found to be surprisingly good (solving 44 of the coiled-coil and 7 of the transmembrane test cases), but the performance of AMPLE was significantly better. AMPLE's truncation and ensembling pipeline was then applied to the solution of protein structures using very distant homologs, and compared with the performance of the current state-of-the-art in automated Molecular Replacement in MRBUMP. The AMPLE pipeline was able to solve structures that could be be solved using MRBUMP, showing how AMPLE is able to find the evolutionarily conserved structural core from homologs that cannot be accessed using existing protocols. Work was also carried out to optimise AMPLE's cluster and truncate procedure. This has resulted in a significant improvement on AMPLE's ability to solve the structures in a difficult set of test cases (solving 11 of 18 test cases compared with 6 for the original protocol), despite only a modest increase in processing time. As part of this work, AMPLE has been extended from a prototype piece of software consisting of a collection of independent scripts, to a coherent, modularised program incorporating a range of software best practice.
|
|---|
