Co-evolving protein sites: their identification using novel, highly-parallel algorithms, and their use in classifying hazardous genetic mutations

Knight, Louise ORCID: https://orcid.org/0000-0003-1431-197X 2017. Co-evolving protein sites: their identification using novel, highly-parallel algorithms, and their use in classifying hazardous genetic mutations. PhD Thesis, Cardiff University. Item availability restricted.

Preview	PDF - Accepted Post-Print Version Download (2MB) | Preview
	PDF - Supplemental Material Restricted to Repository staff only Download (2MB)

Abstract

Algorithms for detecting molecular co-evolution have until now been applied only to individual protein families, but not to the human proteome. Linked to this is the problem that performing the computations for identifying co-evolving sites in the human proteome would take a prohibitively long time using the serial algorithms already in use. In addition, co-evolving sites have not been pursued as a possible way of classifying mutations according to their likelihood to cause disease. The main contributions of this thesis are as follows: identification of three suitable methods for detecting molecular co-evolution by comparing the performance of published state-of-the-art methods on simulated data; implementation of these methods in the parallel architecture CUDA, and evaluation of these methods’ performance in comparison to serial implementations of the same methods; and identification of co-evolving sites across the entire human proteome, and analysis of these sites according to what is already known about disease-causing mutations. Beyond demonstrating the effectiveness of CUDA for implementing molecular co-evolution detection algorithms, we derive insights into techniques for efficient implementation of algorithms in CUDA (particularly algorithms which require tree-based structures, such as parametric methods), and our results provide preliminary insights into the relationship between co-evolving sites and mutation pathogenicity.

Item Type:	Thesis (PhD)
Date Type:	Completion
Status:	Unpublished
Schools:	Computer Science & Informatics
Date of First Compliant Deposit:	7 February 2018
Last Modified:	03 Nov 2022 10:42
URI:	https://orca.cardiff.ac.uk/id/eprint/108951

Actions (repository staff only)

Edit Item