Thesis (Ph. D.)--Hong Kong University of Science and Technology, 2003. / Includes bibliographical references (leaves 118-137). Also available in electronic version. Access restricted to campus users.

Thesis (M. Phil.)--University of Hong Kong, 2010. / Includes bibliographical references (leaves 86-101). Also available in print.

For the last 40 years, the systems community has invested a lot of effort in designing techniques for building fault tolerant distributed systems and services. This effort has produced a massive list of results: the literature describes how to design replication protocols that tolerate a wide range of failures (from simple crashes to malicious "Byzantine" failures) in a wide range of settings (e.g. synchronous or asynchronous communication, with or without stable storage), optimizing various metrics (e.g. number of messages, latency, throughput). These techniques have their roots in ideas, such as the abstraction of State Machine Replication and the Paxos protocol, that were conceived when computing was very different than it is today: computers had a single core; all processing was done using a single thread of control, handling requests sequentially; and a collection of 20 nodes was considered a large distributed system. In the last decade, however, computing has gone through some major paradigm shifts, with the advent of multicore architectures and large cloud infrastructures. This dissertation explains how these profound changes impact the practical usefulness of traditional fault tolerant techniques and proposes new ways to architect these solutions to fit the new paradigms. / text

Not available / text

Previous theoretical work on DNA replication neglected how the starting points (origins) take their place and how replication time is a ected when origins fail to activate. It is however crucial that origin loci are chosen so that too large gaps between them are avoided; otherwise the time until completion of chromosome replication becomes much longer than is allowed by the cell cycle. We investigate what the optimal origin location should be depending on the likelihood of origins failing. We show analytically and numerically that there exist regimes for origins, either to be positioned together in groups spaced far away from the next, or as equally scattered single origins depending on the uncertainty when activation occurs. The model reproduces origin distributions of frog embryos which are thought to be random, and shows contrarily that grouping must occur in order to swiftly complete replication; known as the random completion problem. The model also holds when considering a circular DNA topology for archaeal genomes, as well as if applied to the whole replication pro ling data of yeast. We study how an optimal origin distribution can arise and propose a mechanism to solve the random completion problem. We show that regular spacing emerges as an inherent property of the car parking model. We introduce a spatial requirement for origins to bind to DNA; origins occupy space on the DNA and can only bind stably if there is su cient space for them. Such a model leads to a well ordered origin distribution with minimal gaps as required for on time DNA replication in frog embryos. The optimal origin distribution emerges directly from our model because origins have a higher chance to bind to large empty regions instead of small once, therefore destroying large inter origin gaps. We also introduce a model to account for the interaction of replication forks with each other which leads to their assembly into replication factories. We show using Boltzmann statistics that their assembly is stochastic. A rst model only considers two pairs of forks which we then extend to describe properties of measured experimental distributions such as fork numbers per factory during on a whole yeast genome approach. Our in silico distribution of forks per factory matches in vivo data well; which suggests that active forks encounter each other randomly for an association into replication factories.

No description available.

Replication fork stalling is a source of potentially tumourigenic genomic instability. The RecQ family helicase, Rqh1, is critical for the prevention of replication fork collapse and the formation of potentially deleterious recombination intermediates following fork stalling. Previous work in our lab with Schizosaccharomyces pombe (fission yeast) has shown that rqh10/rqh10 diploids are inherently unstable and show rapid reversion to the haploid state. The current work exploits this characteristic of fission yeast rqh10 mutants in a screen for genes that normally promote replication associated genomic instability. The rqh10rad30 mutant strains employed in this work incorporate the checkpoint deficiency caused by a lack of Rad3, so as to exacerbate the genomically unstable nature of this model. The current work describes the lithium acetate transformation based random mutagenesis by non-homologous integration of the ura4+ selectable marker into the rqh10rad30 fission yeast strains. This random mutagenesis generated extensive (24,500 – 50,000) mutant libraries. The quality of the libraries was assessed by can1 mutant assay, confirming an adequately extensive mutagenesis for the proposed screen. The process to be employed in the screen would involve the crossing of the mutant libraries, with the hope of generating diploids that will have two mutant copies of the same gene. Some of these diploids would appear unusually stable, showing a normal sporulation phenotype. This would indicate the mutation of a gene that normally promotes genomic instability following replication fork stalling. The practicality of the proposed screen of a vast number of diploids was assessed and described in detail in the current work. A technique involving inverse PCR (IPCR) adopted from previous work to identify mutants of interest, was also investigated. The investigation of this technique, and the work of others, suggests that transformation using such selectable marker fragments results in most apparent transformants containing extrachromosomal ura4+ fragments. These fragments are thought to provide the predominant template for IPCR, rendering the process unsuccessful at identifying the mutation in the current screen. However, with the mutant libraries generated, and the screen procedure detailed, the stage is set to conduct the screen once a more appropriate mutation location technique is identified. / Thesis (Master, Pathology & Molecular Medicine) -- Queen's University, 2008-05-31 22:25:14.009

Transient replication assays have identified a late expression factor 3, LEF-3, to be essential for DNA replication and late gene expression in the baculovirus species, AcMNPV. Although its specific role in these two processes has not been determined, this single-stranded DNA binding protein is multi-functional. LEF-3 forms a homo-oligomer, binds single-stranded DNA, interacts with components of the viral replication complex and is required to transport the helicase protein P143 into the nucleus of infected cells. Various regions within LEF-3 were deleted to determine the domain essential to its function and the N-terminal amino acids 1-125 were found to be sufficient for late gene expression. This N-terminal region includes the 56 amino acid region of LEF-3 required for nuclear transport. In order to define this domain, the effect of site-specific mutagenesis of LEF-3 on its intracellular distribution was determined. Fluorescence microscopy of expression plasmid transfected cells demonstrated that amino acids 14 to 37 formed the core nuclear localization signal (NLS), but the flanking amino acids may act as regulatory elements. Comparison with other group 1 Alphabaculoviruses suggested that this core region contained a functionally duplicated NLS. The AcMNPV LEF-3 also functioned in mammalian cells indicating that the protein nuclear import systems in insect and mammalian cells are conserved. Mutagenesis of two conserved cysteine residues located at amino acids 82 and 106 were not essential for nuclear localization or for interaction with P143. However, by using a modified construct of P143 that localized on its own to the nucleus, it was demonstrated that a functional nuclear localization domain on LEF-3 was required for an interaction between LEF-3 and P143. / Thesis (Master, Microbiology & Immunology) -- Queen's University, 2009-01-04 18:25:23.7

No description available.

No description available.