Use this URL to cite or link to this record in EThOS:
Title: Applications of knowledge-tunnelling to crystallographic refinement of macromolecules
Author: Gore, S.
Awarding Body: University of Cambridge
Current Institution: University of Cambridge
Date of Award: 2007
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
The vastness and complexity of the conformational landscapes present a major challenge to structure determination of macromolecules. Procedures using energy functions and statistical preferences modify a conformation in contrasting ways (continuous and discrete respectively).  Knowledge-tunnelling is a discrete knowledge-based modification that can rescue a conformation from a local minimum of an energy function. The complementarity between knowledge-tunnelling and kinetic sampling has been exploited by the RAPPER program for ab initio protein loop prediction and automatic crystallographic refinement of an approximate protein model. But RAPPER’s applicability is limited to sampling proteins only in a way dictated by an algorithm that is hard to modify. This thesis addresses these limitations to extend the promise of knowledge-tunnelling to new challenges in macromolecular crystallographic refinement. RAPPER has been reformulated as Rappertk a modular and scriptable software framework that facilitated the conformational sampling tasks undertaken later. Along with RAPPER’s genetic branch-and-bound algorithm, a combinatorial optimization algorithm for sidechains is implemented in Rappertk. Ligand and RNA sampling are incorporated in addition to restraint-sensitive protein sampling. For crystallographic refinement of proteins, optimal sidechain placement, loop sampling techniques are developed by using the discrete residue specific conformational preferences of mainchain and sidechains.  For RNA refinement, rotameric preferences of sugar-phosphate backbone are used. In all cases, the value added by knowledge tunnelling is assessed by comparing the CNS/Rappertk composite protocol with the CNS-only preferences used, knowledge-tunnelling improves the refinement statistics and is useful in sequence assignment, all-atom refinement and estimation of structural heterogeneity.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available