UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Protein structure prediction and modelling

Swindells, Mark Basil; (1992) Protein structure prediction and modelling. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of Protein structure prediction and modelling.pdf]
Preview
Text
Protein structure prediction and modelling.pdf

Download (23MB) | Preview

Abstract

Empirically based methods are developed for the analysis and prediction of protein structure. All of the approaches are based on information derived from analyses of atomic resolution protein structures. In this thesis, two related projects are described. The first considers secondary structure prediction. An automated procedure is developed which combines new approaches with previously published methods. The important feature of this new algorithm is the development of a reliability index, which assesses the likelihood of each prediction being correct. Rigorous assessments of the algorithm confirm that certain regions of proteins can be predicted with enhanced reliability. This work is followed by studies of the most confident, yet incorrect predictions. Using these results, the limitations of secondary structure prediction are discussed. The second part of this thesis analyses the interleukin-1 family of cytokines and the Kunitz family of trypsin inhibitors. Despite being topologically similar, they have no detectable sequence similarity when using standard alignment procedures. By implementing both sequence and structural analyses of the two families, various similarities are observed. The most important is the maintenance of a tightly packed hydrophobic core which is limited to only 10% of the structure. The constituent side chains are tightly packed in both families. Any mutations that occur are limited to a small set of hydrophobic amino acids, which compensate in such a way that the core volume remains similar. These analyses suggest that the maintenance of a hydrophobic core is crucial for determining this fold and that any methods developed for fold recognition will need to take this into consideration. Modelling experiments are subsequently performed on the recently cloned interleukin-1 receptor antagonist, in order to identify surface regions which may be responsible for receptor binding and signal transduction. These are of particular interest as elevated interleukin-1 production may be implicated in rheumatoid arthritis.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Protein structure prediction and modelling
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Thesis digitised by ProQuest
Keywords: Pure sciences; Protein structure
URI: https://discovery.ucl.ac.uk/id/eprint/10111814
Downloads since deposit
114Downloads
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item