Use this URL to cite or link to this record in EThOS:
Title: New computational methods for the design of peptidomimetics
Author: Garland, S. L.
Awarding Body: University of Cambridge
Current Institution: University of Cambridge
Date of Award: 1997
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
Many peptides confer biological activity at a receptor through the interaction of a very small number of side-chains, typically 2-4, arranged in a conformation based on the β-turn. The design of novel peptidomimetics has been simplified on this basis; scaffolds were selected to mimic fragments of the β-turn rather than the structure as a whole. Cluster analysis of Ca-atom positions provided distance constraints that typify potential mimetics and may be used for efficient screening of chemical structure databases. A similar analysis of the Ca-Cβ bond vectors was performed for the 11 well-defined, idealized, β-turn types. All 2 and 3 side-chain fragments (2,640 in total) fall into just 10 and 12 significant classes, respectively. These classes provide model constraints for the design of β-turn peptidomimetic libraries, the design of one structure per class optimally spanning the conformational space available to β-turn fragments. De novo structure generation methods were applied to the simplified design constraints. Thorough analysis of the results required structure classification procedures, two of which have been developed and assessed. The method of choice was found to be the SMILES code, which provides a unique string representation of molecular connectivity. This has been used to monitor the structure generation process. The rate of discovery of structures with novel connectivity decreases over time and extrapolation to the asymptote provides an estimate of the total number of solutions possible for a given set of design constraints. The vast majority (estimated at over 95%) of possible 2 and 3 side-chain β-turn peptidomimetic scaffolds have been designed, providing many thousands of diverse novel structures. Subsequent classification has revealed certain structures that can mimic more than one class of β-turn fragment through changes in conformation. These results are likely to be important in the traditional design of peptidomimetic combinatorial libraries for screening.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available