Investigation of an immune algorithm and differential evolution to study folding of model proteins

Bennett, Andrew James (2010). Investigation of an immune algorithm and differential evolution to study folding of model proteins. University of Birmingham. Ph.D.


Download (12MB)


The application of computational search techniques to global optimisation problems is becoming increasingly popular. Search techniques have been used to study the folding of model proteins, with the aim of accurately predicting the native state of a protein from its amino acid sequence. Through modelling, knowledge of the folding process can be obtained. In this thesis, two search techniques have been applied to a variety of protein models. The development and application of both an Immune algorithm and a Differential Evolution search technique are described, with the aim of finding the lowest energy conformations of coarse-grained, model proteins. Initially, the two-dimensional HP Lattice Bead Model is investigated, followed by three-dimensional models of varying complexity. The HP Lattice Bead and BLN models, on a diamond lattice are considered, as well as the Dynamic lattice Model, using backbone torsion angles to define the structure of the lattice. A modified chain growth constructor is introduced; firstly, to generate the initial population for both search techniques, secondly, to record unoccupied lattice sites of meta-stable conformations to reduce the risk of performing infeasible joint mutations during the mutation phase for the Immune algorithm, and thirdly, to improve the standard of mutations performed by Differential Evolution. A novel profiling system is introduced based on the theory of genealogy and ancestry by recording the parent of each individual. The method is used to track and evaluate the diversity of populations and assess the impact that genetic operators have on this diversity. The aim of applying this system is three fold: to investigate how effective genetic operators are; to allow a greater understanding of the progress of the optimisations; and to assess the strengths and weaknesses of each search technique investigated.

Type of Work: Thesis (Doctorates > Ph.D.)
Award Type: Doctorates > Ph.D.
College/Faculty: Colleges (2008 onwards) > College of Engineering & Physical Sciences
School or Department: School of Chemistry
Funders: None/not applicable
Subjects: Q Science > QD Chemistry


Request a Correction Request a Correction
View Item View Item


Downloads per month over past year