Search

The RSKs are known to be involved in many cancers, including Prostate Cancer (PCa), however the representation of the Ribosomal S6 Kinase (RSK) isoforms is skewed and there is little investigation into the RSK regulation of the Androgen Receptor (AR). Here we show in model African Green Monkey Kidney Fibroblast (COS-1 cells) that all RSKs regulate the AR transcriptional activity and interact with the AR in proximity. Analysis of PCa cell lines highlighted that all RSKs exhibited higher expression than the control but most significantly RSK2. We then transitioned to lymph node carcinoma of the prostate (LNCaP) cells as they are more biologically relevant, siRNA knockdown coupled with qPCR showed that key AR target genes was differentially regulated by the RSKs; RSK1 and RSK4 knockdown caused increased expression of NDRG1; RSK1 and RSK2 knockdown induced a decrease in TMPRSS2; RSK2 knockdown produced a significant decrease in PSA, respectively. To understand if these effects would translate to a cellular change we investigated proliferation, siRNA knockdown of RSK2 and RSK4 caused a significant decrease in LNCaP cells. Interestingly, investigation of RSK4 is also underrepresented in structural biology despite having a different mechanism from other isoforms. Here, we attempted to express, purify, and crystalize different domains of RSK4 to understand its mechanism of activation. Crystallization attempts were unsuccessful, potentially due proteolysis and chaperone contaminations. Kinetic assays showed that phospho-mimetic mutations were successful in creating a constitutively active mutant construct. Pull down assays experimentally showed that the AGC region of RSK4 is essential for the PDK1 independent activation. The multiple techniques used in this study strongly suggest that RSK are in fact involved in PCa signalling and warrant further investigation and highlight the nonredundant functions of the RSK isoforms and a novel relationship between AR and RSK4. Additionally, the pull-down assay highlighted that the AGC region of RSK facilitates interaction between the N- And C-terminal domains which has not been previously experimentally demonstrated thus supporting previous speculation around its activation mechanism.

In railways, turnouts, or also called switches and crossings, are widely reported as the origin of many problems owing to their involvement in many accidents and the major share on the budget of infrastructure managers. Turnouts have an asymmetrical and large geometry in order to enable diversion between railway tracks, leading to complex dynamic behaviour of the structure. The complex structure of turnouts and its dynamic effect catalyse the process of many failure types. One of common failures over a turnout structure is dynamic fragility. Particularly, high-frequency high-magnitude impact forces that are observed over turnout crossing noses cause many dynamic fragility problems including long terms defects such as fatigue. This thesis aims at establishing a new numerical method to simulate dynamic fragility of turnout crossings from fracture mechanics' point of view. In order to achieve this aim, first, it is targeted to improve the understanding of dynamic behaviours of turnouts under impact loadings varying from ideal environments to extreme events. A new numerical methodology is proposed to evaluate the complex behaviours of turnouts and its components. The new insights stemmed from this research has shown the importance of taking dynamic effects into account and the force spectra observed over the turnout structure. More importantly, the Multiphysics informed models helped to determine the critical section over a turnout structure and corresponding failure modes. Following having a more comprehensive understanding of dynamic behaviours of railway turnouts, the fatigue behaviour of the critical section has been examined via Peridynamics. Peridynamics is a recent and promising numerical tool in the field of Fracture Mechanics. Prior to fatigue investigation, two new methodologies have been introduced into Peridynamics theory. Those methodologies have helped to track the formation of cracks, and reduce the simulation time significantly. It would not be possible to conduct fatigue simulations of complex geometries without those contributions. As is the final target, fatigue simulations have been carried over critical sections of a turnout as well as experimental geometries. It was found that the fatigue model in Peridynamics is promising in certain damage modes and confirms the general knowledge in most cases. Nonetheless, its efficiency has been found questionable in the case of fatigue over the crossing nose under a mixed mode compression loadings. The characteristics of Peridynamics have imposed certain limitations and it seems that those challenges must be addressed first in order to be able to increase the efficiency and to apply in railway applications.

Projects can be a collection of numerous interdependent subsystems, including activities, resources, information, etc. Complexity has been found to significantly impact the realisation of projects in terms of scope, time, and cost, and hence influences the discipline of project management. Studies confirm that most megaprojects do not meet their defined scope, time, and cost estimations and hence can be considered project failures. Despite many research studies considering these issues, projects still fail. This has been primarily due to a poor understanding of complexity, particularly the non-linear and dynamic interactions and interdependencies between project elements and the project and its surrounding environment. These issues can lead to deviations from defined objectives, delays, and cost overruns. Nonlinearity in complex projects causes unpredictability in the relationship between inputs and outputs. The behaviour of nonlinear systems can, on occasion, be predicted by employing some qualitative patterns, but complex systems are generally not responsive to conventional systems analysis. As modern projects are becoming more complex, new types of management approaches for projects are needed. Research is therefore needed to both: (i) understand complex issues within projects and (ii) decide how to manage complexity appropriately. Railway infrastructure projects are particularly vulnerable to disruptive events caused by uncertainties and internal and external influences (such as scope change, change in design etc.) during their lifecycle. The adaptive capacity of systems refers to either the ability of the system to (i) return to its equilibrium point when disruption is encountered, using its programmed strategies, and planning methods, or (ii) adopt new approaches to respond to events that are outside of its preconfigured design and structure. The measure of the rapidness of a project to recover from disruption is referred to as resilience. The literature review highlighted a few of future research avenues to achieve a reliable tool to assess project resilience : (i) apply methods in more advanced fields (i.e. applicable to megaprojects); (ii) identify a set of validated indicators that can assess a project's resilience, and hence its ability to manage disruptions; (iii) conduct structured experimental studies based on identified indicators that can evaluate proposed conceptual frameworks for assessing project resilience, and (iv) develop tools to be used by project stakeholders and managers to evaluate the impact of efforts required to enhance the resilience of the existing and future projects. Such a tool could be applied to measure a project's strengths and weaknesses and suggest action plans to improve its resilience. The tool would augment existing project risk management strategies and hence help projects to become more resilient when facing disruptive events or unexpected changes. The scope of the research is to develop and apply an innovative methodology and analysis tool to railway project management. The approach considers project management as a system and utilises concepts and methods of system dynamics, to model the system (and its subsystems) and to evaluate its resilience. To evaluate the real-life application of the proposed method, the author applied his proposed method to measure the resilience of metro systems in response to disruptions or additional operating hours. The results are presented in Chapter 8. The research aims to propose an innovative systematic, and reliable methodology to consider railway project management structure as a complex system and model its subsystems via Causal Loop Diagrams. To do so, the following objectives are defined: • Generate a set of systematically developed causal loop diagrams for key subsystems of project management and conduct qualitative analysis to identify the resilience factors of each subsystem. • Integrate and synthesise the modelled subsystems through developing innovative structured and formulated stock and flow diagrams (SFD) and studying the resilience of the proposed schematic project management structure. Qualitative analysis will be utilised to study interactions between developed subsystems • Convert and propose the developed SFD into an innovative modelling tool, which can be adapted to any railway project and enable users to analyse trade-offs between different variables affecting project management resilience, as well as evaluate the impact of decisions on project performance. • Develop a set of novel and unique CLDs and SFDs to model metro system operation management environment and analyse its resilience to be converted into 24-hour metro operation. The proposed approach and platform of thinking are proposed to be applied as a complementary tool to measure the resilience of project management for railway infrastructure schemes and to add value to the existing project risk management approaches. Project management was considered as a system with five main subsystems, namely, project governance, requirements management, configuration management, in-house engineering and change management. Multiple case study analysis was used to diagnose and identify the main components of the proposed subsystems that most affect the vulnerability and resilience of the system the most. To achieve this, questionnaires were designed to survey the key components and factors, to be finally used as the main variables for generating causal loop diagrams for each subsystem. Then CLDs were quantified to facilitate quantitative analysis in addition to the qualitative analysis. The full methodology is presented in Figure ‎1.5. This thesis presents an innovative combination of qualitative and quantitative analysis approaches founded on system dynamics tools, to evaluate the resilience of the proposed schematic project management structure for railway. The main idea was to propose an innovative and practical approach to applying system thinking to conventional project management strategies to reinforce it and bridge the existing gaps to manage the growing complexity of modern projects., as highly recommended by previous researchers. The proposed schematic project management structure and subsystems are modelled using the data derived from multiple case study analysis, causal loop diagrams (CLDs), and Vensim software. Generated CLDs were reviewed by a group of external experts and some from the studied projects to validate the realism of the CLDs. CLDs are simplified systematically to keep the reliability and at the same time applicability for analysis. CLDs provide a set of cross-validation tools such as the Uses tree and Cause tree, which deliver traceability features of the generated CLDs. Uses trees can be used to trace the impact of any selected variable on the other components of the system., whilst the "Cause Tree" can be used to figure out the variables with an impact on any selected component. Employing these tools, provided a reliable qualitative analysis tool and allowed the author to bring out the causal feedback of each CLD and identify the key resilience indicators for each subsystem of the model (Qualitative analysis). The qualitative analysis of CLDs highlighted the key resilience indicators affecting each subsystem and reflected the causal feedback of the loops. The traceability feature of the CLDs can be adopted as an applicable method to simulate the environment of complex projects and assist managers to visually trace the impact of their decisions on the performance of the project. The research contribution can be briefly summarised as: • Developed and proposed an innovative methodology combining multiple case study analysis and system dynamics tools, which can be followed by future researchers to visualise the complex environment of projects. • Generated a set of reference CLDs to model the railway project management as a system. These models are developed for the first time and deliver a holistic overview of any similar railway project. Researchers can adapt these reference models and map their individual systems against the generated CLDs. Hence, the CLDs can be modified to reflect the true nature of any specific system (Qualitative analysis tool for complex railway projects). • Designed, formulated, and proposed a novel SFD to analyse the resilience of the railway's proposed schematic project management structure, according to the resilience indicators identified via CLD analysis. The author designed a unique user interface for system analysis. This is the first-of-a-kind toolset, which allows managers and researchers to change the resilience indicators of the project management and analyse its impact on project performance and management resilience (A novel quantitative analysis tool applicable to all railway projects). • Based on the findings from qualitative and quantitative analysis, the author developed a set of novel formulas and a conceptual two-dimensional model to enhance the understanding of project resilience. This creates a platform of thinking and proposes new research avenues for future researchers to focus on other dimensions with potential impact on resilience and achieve the optimum level of performance. • Designed and proposed a unique system dynamic tool, formulated, and structured to assist managers and planners to analyse the resilience of metro operation environments. The tools can be sued to manage disruption in the metro system or in the case of converting metros to 24-hour metro operation and analyse its impact on metro system resilience.

Atmospheric CO₂ (Ca) is rapidly increasing and expected to reach, and perhaps surpass, over 550 μmol mol 1 by the middle of the 21st century. Rising Ca is the principal cause of human-induced global climate change and is primarily the result of anthropogenic activities such as fossil fuel burning and irreversible land-use change. However, the ocean and many terrestrial ecosystems take up and store carbon (C), therefore not all emitted CO₂ stays in the atmosphere. The terrestrial biosphere has an important role in the C cycle yet there are large uncertainties in its response to future environmental change. Forests are vital in the C cycle as they are large carbon stores with very large fluxes to and from the atmosphere. Forests are, therefore, important when considering the future atmospheric composition. In the northern hemisphere, temperate deciduous forests are considered to be particularly important carbon sinks yet there are still knowledge gaps regarding the responses of mature temperate forest ecosystems to elevated Ca and how this will affect the functioning of plants and ecosystems. Free-Air Carbon Enrichment (FACE) facilities enable us to assess and quantify the effects of elevated CO₂ (eCO₂) on whole patches of forest systems. This thesis aims to enhance the knowledge base for the effects of eCO₂ on the leaf-level physiology, biochemistry, and morphology of mature Quercus robur using a Free-Air Carbon Enrichment (FACE) facility at the Birmingham Institute of Forest Research (BIFoR). The objectives of this study are to quantify the photosynthetic performance, foliar nitrogen, and chlorophyll content, in oak leaves under simulated future (i.e., elevated) CO₂ conditions. Additionally, a meta-analysis, including new data on Q. robur, was conducted to examine the stomatal responses across eCO₂ experiments from 1999 to 2019 and assess the use of the optimal stomatal theory principles to characterise stomatal behaviour with eCO₂. This study, using tree canopy measurements, found that net photosynthesis (Anet) increased by 24% in mature Q. robur under eCO₂ (150 mol mol 1 above Ca) and that photosynthetic enhancement was sustained across the first three years of CO₂ fumigation at BIFoR FACE. There were no changes to foliar nitrogen (N) or chlorophyll under eCO₂ exposure, but season and canopy position were found to be important influential factors when scaling leaf-to-ecosystem light harvesting and N allocation. The meta-analysis reported increases in intrinsic water use efficiency (iWUE) that were similar across the plant functional types (PFTs) that were assessed. There was no significant difference in stomatal responsiveness among the PFTs and no significant reductions in stomatal conductance (gs) with eCO₂. Additionally, the optimal stomatal theory successfully characterised stomatal behaviour with eCO₂ across a range of environmental conditions. The research presented in this thesis extends the current knowledge on the physiological, biochemical and morphological responses of mature forest-grown trees in response to eCO₂ and identifies the key environmental factors that influence the eCO₂ response. The information presented contributes significant new knowledge that may be used by vegetation modellers (for the calculation of future carbon budgets) to help guide government policy developers and forest managers in decision making regarding climate change mitigation and to increase forest and environmental resilience to the effects of climate change such as with eCO₂.

Background and Aim: Atrial fibrillation (AF), the most common sustained cardiac arrhythmia, is a frequent symptom of inheritable cardiac arrhythmia syndromes. The clinical picture of AF develops due to a multitude of mechanisms which are aggravated by external factors and comorbidities, complicating explanatory research. Hence, studying inheritable drivers of atrial arrhythmia represents a valuable approach to pinpoint molecular pathomechanisms of AF. Aim of this thesis therefore was to gain knowledge on atrial arrhythmia mechanisms driven by genetic predisposition. Additionally, we set out to investigate the impact of elevated androgenic steroid concentrations as an extrinsic factor. Methods: Clinical observations were used as a basis for murine experimental study designs. Employing two murine models, one for arrhythmogenic right ventricular cardiomyopathy (ARVC) and one for familial AF, enabled us to study atrial arrhythmia mechanisms at different scale, from in vivo to ex vivo, from the living animal to ion channels in atrial cardiomyocytes. Effects of the mutations and interventions on atrial structure, function and electrophysiology were revealed by a variety of methods, including echocardiography, electrocardiograms (ECGs), optical mapping, monophasic action potential recordings, histology, RNA sequencing and super-resolution microscopy. Moreover, atrial involvement in human ARVC was studied in digital patient ECGs. Results and Conclusions: Atrial arrhythmias are a common clinical feature of ARVC and are accompanied by impaired conduction, observed in both, patients and the murine model. Vulnerable desmosomes and increased concentrations of androgenic steroids jointly impair atrial conduction and cardiac sodium current. The studied point mutation of the cardiac sodium channel alpha-subunit Nav1.5, associated with familial AF, leads to a gain-of-function and atria show a blunted response to the clinically used sodium channel blocker flecainide. Our findings underline that murine models represent a valuable tool to decipher drivers of atrial arrhythmia seen in patients.

Lung cancer is the leading cause of cancer death worldwide with curative intent surgery the most effective treatment. However, there is a risk of developing postoperative pulmonary complications (PPC) including pneumonia and clinically significant lung atelectasis. This thesis investigated the role of cigarette smoking in PPC development, including the impact of smoking cessation and the role of key immune cells in the lung, alveolar macrophage (AMs), with the aim to develop strategies to prevent PPCs in those who smoke before major lung surgery. In a retrospective study of prospectively gathered data from a cohort of patients undergoing major lung surgery between April 2010 and April 2014, the risk factors and short and long-term outcomes of PPC development were investigated. In those patients with lung cancer diagnosis undergoing major lung surgery, the effect of self-reported smoking status on PPC development and short and long-term outcome after surgery were also investigated. In a translational cohort of patients undergoing lung resection for cancer recruited between August 2018 and August 2021, non-cancerous lung tissue was obtained to isolate AMs and smoking status determined using self-reported status and exhaled carbon monoxide (CO) measurements. AM phenotype was determined and function including efferocytosis, which is a process by which apoptotic neutrophils are removed. A protocol was also developed for a feasibility study of smoking cessation in the lung surgical pathway. Development of a PPC after lung surgery was found to be associated with worse short-term and long-term outcomes after major thoracic surgery. Cigarette smoking was the biggest risk factor for PPC development, with current smokers having the highest risk of developing a PPC. Self-reported status alone, using a 6-week cut off to determine recent and long-term ex-smokers, the optimum timing for smoking cessation before lung surgery could not be defined. In the translational cohort of patients, the use of exhaled CO found that 1 in 5 patients continued to smoke immediately before surgery. AMs from smokers had increased expression of cell surface markers involved in inhibition of efferocytosis (SIRPα and CD33), and reduced expression of cell surface markers involved in promoting efferocytosis (CD206 and CD163) compared to long-term ex-smokers and never smokers. AMs from patients who were smoking had higher levels of SIRPα, with lower levels found in those who stopped after 4 weeks. Furthermore, AM SIRPα levels were higher and efferocytosis function reduced in those patients who developed PPC. Smoking is the biggest risk factor for PPC development after major lung surgery, the feasibility of an integrated smoking cessation programme including the web-based application Quit4Surgery needs to be assessed. Modifying SIRPα has a potential role in restoring AM function in those who continue smoking before surgery: the mechanisms of cigarette smoking on SIRPα expression are an area for future research.

The Ikaros Family of transcription factors Aiolos, Eos, Helios, Ikaros and Pegasus are a family of zinc finger proteins involved in the regulation of genes within a plethora of immune cells. Studies have suggested the family play a role in mediating Regulatory T cell (Treg) and T helper 17 (Th17) responses within Inflammatory Bowel Disease (IBD) but much of the detail of their function and expression is undocumented. This thesis examines the expression and roles of the family in IBD and the associated autoimmune liver disease Primary Sclerosing Cholangitis (PSC). Using qPCR and immunohistochemistry, we investigated the expression of the Ikaros family in blood derived CD4+ T Cells, myeloid cells, and whole bowel and liver tissue from diseased and healthy individuals. I also used an siRNA knockdown of Eos (siEos) within THP- 1 derived macrophages to investigate changes in phenotype and function. We report altered expression of family members within diseased bowel and liver in CD4+ T cells, but also myeloid cells. This includes previously unreported expression of the family within macrophage populations, with changes observed in macrophage subtypes. Ikaros family members were present in both the bowel and liver with greater numbers of Eos and Helios-expressing macrophages found in diseased bowel. Importantly, Eos expression was elevated in monocytes from both Biologic Naïve and anti-TNFalpha treated patient samples in comparison to Control, with novel correlations between Ikaros family members and liver clinical markers also identified. Using siRNA, we found siEos macrophages have reduced phagocytic capacity and upregulated proinflammatory mediator production suggesting Eos promotes a regulatory macrophage phenotype. Together, our results suggest a novel role for the Ikaros family within macrophage phenotype in bowel and liver disease.

Osteoarthritis (OA) is one of the considerable chronic health conditions worldwide. In the UK, 8.75 million people were receiving OA treatment. Currently, there are no disease-modifying drugs, and treatment relies on the management of pain associated with OA. The delivery of OA drugs is challenging because the cartilage is dense, avascular (i.e., has no blood vessels), and relies on diffusion to transport nutrients and drugs to the cells. Moreover, cartilage matrix is composed of anionic proteoglycans that can repel negatively charged drugs, making it difficult for OA drugs to penetrate through the matrix and reach the target cells. Additionally, most OA drugs are hydrophobic, so an appropriate technique is required to deliver them inside cartilage. A novel and efficient emulsion-based local delivery of OA drugs coated with poly beta-amino ester polymers (POLY-X) into cartilage was proposed. In this project, nonsteroidal anti-inflammatory drugs as a model drug for OA treatment were examined. The characterization of the developed delivery system and its efficacy in native (nontreated) and glycosaminoglycan (GAG) depleted cartilage was assessed by measuring the zeta potential and size, amount of drug uptake and retention in the cartilage. In addition, the ability of the developed drug delivery system to inhibit cytokine (IL-1α) induced glycosaminoglycan and collagen degradation, as well as the loss of cell viability in cartilage explants were tested. The data showed that the developed emulsion delivery system exhibited enhanced and prolonged drug localisation not only on non-treated cartilage tissues but also on GAG depleted sample resulting in a higher amount of drug uptake and retention in cartilage compared to the control, and the difference was statistically significant (p < 0.05). Furthermore, the developed emulsion delivery system protects the viability of the chondrocyte cell and provides a significant increased (p <0.01) in glycosaminoglycan and collagen content compared to IL-1α treated cartilage and similar to untreated control. The percentage of sGAG and collagen loss was 3 to 5 times higher in IL-1α treated cartilage compared to the untreated and developed drug delivery system treated groups. This study proved that modified emulsion-based therapy could provide a substantial improvement in the treatment of OA to maintain cartilage properties and improve OA outcomes.

Zeb1 is a zinc finger transcription factor that plays a critical role in embryonic development, tissue regeneration, and stem cell biology. While Zeb1 has been suggested to be involved in regulating the self-renewal and differentiation of hematopoietic stem cells (HSCs), the cellular and molecular mechanisms by which Zeb1 may regulate HSC function are not fully understood. We initially identified potential ZEB1 target genes in HSCs by RNA-sequencing of adult HSCs after Mx1-Cre mediated deletion of Zeb1. We found Zeb1 regulates cell adhesion and cytoskeletal organization related genes as well as lipid metabolism related genes. Our studies identified cell adhesion molecule EpCAM as a critical ZEB1 target in HSCs. We analysed the impact of Zeb1 mediated regulation of EpCAM expression on the transcriptional programming and molecular fate of HSCs and found that EpCAM expression in Zeb1 knock-out (KO) HSCs increased cell survival and decreased transcriptional markers for mitochondrial metabolism, RNA synthesis, and differentiation. To investigate the role of Zeb1 in the long-term maintenance of HSCs, we conditionally deleted Zeb1 using the Vav-iCre system, which deletes Zeb1 gene expression in HSCs during embryonic development at E11.5. Zeb1 KO mice were born with reduced Mendelian inheritance, demonstrating Zeb1 KO in HSCs during embryonic development causes perinatal lethality. We found Zeb1 KO mice that survived into young adulthood (8-12 weeks) displayed expanded HSC numbers but with reduced repopulation activity in transplantation experiments. Young adult Zeb1 KO HSCs displayed widespread myeloid and lymphoid differentiation defects including B-cell maturation and erythroid differentiation. Disrupted erythroid differentiation in Zeb1 KO HSCs led to the development of anaemia. In addition to the well-described critical role for Zeb1 in T-cell development, we found that loss of Zeb1 in HSCs impaired positive selection of thymocytes and the cell survival of T-cell subpopulations, resulting in a pronounced expansion in effector and memory T-cell frequency and a reduction in naive T-cells. Our RNA-seq analysis of young adult Zeb1 KO HSCs revealed altered transcriptional programming related to erythropoiesis, immunodeficiency, HSC maintenance pathways and inflammation. Heightened expression of v inflammatory cytokines IL-6, TNF-alpha and IFN-gamma in the plasma of young adult Zeb1 KO mice was observed, identifying Zeb1 as a critical mediator of inflammation. Ageing of adult Zeb1 KO mice lead to fatality with almost complete penetrance, and evidence of profound anaemia and preliminary evidence suggesting the development of pre-leukaemia (MDS/MPN-like) was observed. In addition, a single aged Zeb1 KO mouse appeared to develop a CD4+ T-cell driven lymphoproliferative syndrome. This suggests that Zeb1 depletion in HSCs can drive both myeloid and lymphoid malignancies/pre-malignancies. Overall, these data support the notion that Zeb1 acts as a tumour suppressor in select haematologic malignancies. In summary, this study reveals that Zeb1 is critical regulator of long-term HSC integrity, and that lack of Zeb1 in HSCs leads to the development of haematologic malignancies. These findings shed light on the complex molecular mechanisms underlying Zeb1 regulation of HSC function, which may have important implications for the development of novel therapies for hematopoietic disorders where ZEB1 expression is downregulated.

Background: Atherosclerosis continues to be a major contributor to cardiovascular disease (CVD), which is one of the leading causes of morbidity and mortality globally. The current pharmacological strategies targeting hyperlipidaemia, such as statins, have shown limited effectiveness in combating cardiovascular risk and have other issues. Considering the limitations associated with statins and other pharmacotherapies, alternative avenues need to be sought. Nutraceuticals, such as resveratrol (RSV), have been highlighted as potential candidates for atherosclerosis prevention and treatment due to its demonstrated ability to modify several atherogenic risk factors and its excellent safety profile. Unfortunately, its effects on the full range of atherosclerotic processes along with the underlying molecular mechanisms are not fully understood. Therefore, the main aims of this study were to investigate the effects of RSV on key cellular processes associated with atherosclerosis development in vitro and to elucidate its effects on atherosclerosis progression in a mouse model system. Methods: Various in vitro assays were carried out using different cell lines and primary cell cultures to investigate the effect of RSV treatment on a range of key cellular processes associated with atherosclerosis development. Furthermore, to investigate the effect of RSV on atherosclerotic plaque progression in vivo, 8-week-old male low-density lipoprotein receptor-deficient (LDLR-/-) mice were fed either a high-fat diet (HFD) or HFD-supplemented with RSV for 12 weeks. This was followed by a comprehensive analysis of risk factors associated with disease initiation and progression, such as plasma lipid profile and staining of resident cells (e.g. macrophages, T-cells and smooth muscle cells (SMCs)) in the plaque. Results: RSV attenuated several key atherosclerosis-associated processes in vitro, such as monocyte migration towards monocyte chemoattractant protein-1(MCP-1), reactive oxygen species (ROS) production in all investigated cell types, and foam cell formation. Furthermore, RSV reduced human aortic smooth muscle cells (HASMCs) invasion and enhanced their proliferation, and exhibited anti-inflammatory actions. Regarding in vivo progression study, mice that received RSV-supplemented HFD for 12 weeks showed an improvement in plasma lipid profile, attenuation of plaque inflammation and enhancement markers of plaque stability. Furthermore, additional investigation on liver samples showed that RSV has the ability to reduce steatosis. Conclusion: The findings from this study provide valuable insights into the anti-atherogenic actions of RSV and implicate it as a potential nutraceutical candidate that could be used globally as a part of ongoing atherosclerotic CVD prevention and management strategies due to the lack of undesirable side effects and the comparatively low cost compared to standard pharmacological medications. The potential of RSV should be investigated however in large clinical trials.

The neurons that compose the mammalian cerebral cortex are born from a seemingly uniform population of progenitor cells through a process called "cortical neurogenesis". This process is strictly regulated to ensure progenitors produce neurons at the right time and in correct numbers, and defects in neurogenesis are linked to neurological disorders. Mutations in the X-linked gene protocadherin-19 (PCDH19) lead to PCDH19-epilepsy, a disorder that causes early onset epilepsy and cognitive impairment. Although the disease mechanism is unknown, PCDH19 tissue mosaicism is thought to be a crucial underlying factor. Recent work at the Isabel Martinez-Garay (IMG) lab has suggested that the presence of PCDH19 in neural progenitor cells may be important for regulating cortical neurogenesis. Progenitors that expressed either the wildtype (WT) or knockout (KO) PCDH19 allele showed delayed and premature neurogenesis, respectively. Interestingly, this only occurred in the developing brain of heterozygous (HET) mice, since homozygous WT and KO animals showed normal neurogenesis rates. This thesis aimed to examine the molecular mechanisms by which PCDH19 may regulate cortical neurogenesis. To identify molecular hallmarks by which PCDH19 may regulate neurogenesis, bulk RNA-seq and single cell (sc)RNA-seq were performed on cortical tissue from WT, HET, and KO embryonic day (E)11 animals. The bulk RNA-seq initially found few differentially expressed (DE) genes between mutant and WT same sex counterparts. However, enrichment analyses suggested several pathways were upregulated in HET and KO animals, including pathways related to ribosomes, BMP signalling, and Wnt signalling. The results from the scRNA-seq analysis revealed numerous DE genes between mutants and WTs when examined within cell clusters of interest (progenitors and excitatory neurons), including genes related to neurogenesis, Wnt signalling, and Shh signalling. Interestingly, HETs also displayed a marked reduction in ribosomal gene expression compared to WTs and KOs. However, one of the most striking findings from this analysis was that despite WT and KO cells within the HET animals displaying altered neurogenesis rates, no DE genes were uncovered between these cells when compared to each other, suggesting they are transcriptionally similar. Data was also generated that revealed a role of PCDH19 in negatively modulating Wnt signalling, potentially through an interaction with β-catenin and the noncanonical Wnt receptor, RYK. Moreover, immunostaining using pLRP6 antibodies revealed no differences in the proportion of pLRP6 cells in HETs, however KO animals showed a slightly increased proportion. Taken together, this thesis provides the first examination of the molecular mechanisms of PCDH19 in cortical development. The transcriptional data generated will help direct future work to unravel the role of PCDH19 as a regulator of neurogenesis.

Previous studies have demonstrated an extracellular cation-sensor, the calcium-sensing receptor (CaSR), plays a central role in lung development, tissue morphogenesis and inflammation in the lung as well as tissue remodelling in other organs. However, the role of the CaSR in lung remodelling is currently unknown. During my PhD I have 1) developed novel methods for determining airway and interstitial lung remodelling in vivo; 2) investigated the role of the CaSR in age-related lung remodelling in vivo using aging mice with targeted CaSR-deletion from SM22α-positive cells; 3) investigated the role of the CaSR in pre-clinical models of IgE/Th2 asthma, alarmin-driven asthma and COPD-like neutrophilic exacerbation in the presence or absence of inhaled negative allosteric modulators (NAM) of the CaSR; and 4) investigated the role of the CaSR in cell signalling and remodelling processes in vitro using human lung fibroblasts exposed to a remodelling mediator, TGF-β, in the presence of absence of NAM treatment. The key findings of this thesis are: 1. The CaSR in SM22α-positive cells is central to age-related airway and interstitial ECM remodelling. 2. Inhaled NAM treatment ameliorates airway remodelling, particularly goblet cell metaplasia, in models of IgE/Th2 asthma and interstitial remodelling and/or inflammation in alarmin-driven asthma. 3. The CaSR is expressed in human lung fibroblasts and NAM treatment ameliorates TGF-β1-induced upregulation of CaSR and associated G-protein expression; non-canonical TGF-β signalling; and cellular processes implicated in lung remodelling, including cell-adhesion, apoptosis, secretion, growth, proliferation and extracellular matrix (ECM) remodelling. Together these findings strongly indicate that the CaSR plays a central role in physiological and pathophysiological lung remodelling.

Upper-limb prostheses are either too expensive for many consumers or exhibit a greatly simplified choice of actions, this research aims to enable an improvement in the quality of life for recipients of these devices. Previous attempts at determining the hand shapes performed during activities of daily living (ADL) provide a limited range of tasks studied and data recorded. To avoid these limitations, motion capture systems and machine learning techniques have been utilised throughout this study. A portable motion capture system created, utilising a Leap Motion controller (LMC), has captured natural hand motions during modern ADL. Furthering the use of these data, a method applying optimisation techniques alongside a musculoskeletal model of the hand is proposed for predicting muscle excitations from kinematic data. The LMC was also employed in a device (AirGo) created to measure joint angles, aiming to provide an improvement to joint angle measurements in hand clinics. Hand movements for 22 participants were recorded during ADL over 111 hours and 20 minutes - providing a taxonomy of 40 and 24 hand shapes for the left and right hands, respectively. The predicted muscle excitations produced joint angles with an average correlation of 0.58 to those of the desired hand shapes. AirGo has been successfully employed within a hand therapy clinic to measure digit angles of 11 patients. A taxonomy of the hand shapes used in modern ADL is presented, highlighting the hand shapes currently more appropriate to consider during upper-limb prostheses development. A method for predicting the muscle excitations of the hand from kinematic data is introduced, implemented with data collected during ADL. AirGo offered improved repeatability over traditional devices used for such measurements with greater ease of use.

The tactile system is highly complex. The properties of its central and peripheral components determine the way external stimuli are transformed into perception. At the very first stage, first-order tactile neurons respond to skin mechanical deformation and their activation convey a representation of the sensed object (i.e., encoding). However, there are several open questions regarding which factors can significantly influence the peripheral neural response and hence, perception. The goal of the work presented in this thesis is to provide new evidence about the link between skin properties, object's characteristics, first-order tactile neurons response and discriminative judgments. Chapter One provides an overview of the tactile system with a focus on the peripheral components (e.g., skin, first-order tactile neurons), as well as a summary of the relevant behavioural findings on tactile perception in Young and Elderly. Chapter Two outlines the methods used in this work including psychophysics, a device to present tactile stimuli in a controlled fashion, skin measurement techniques, and manufacturing of fine-textured stimuli. Chapter Three provides an in-depth review of computational models that simulate the response of first-order neurons and how they can be applied for psychophysical research. Chapter Four is the first empirical chapter that evaluates the effects of skin and mechanoreceptive afferent properties on spatial tactile sensitivity in young and elderly participants assessed with the 2-point discrimination task. Chapter 5 is the second empirical chapter that investigates the effects of the interaction between finger and surface properties on the detection sensitivity for a single microdot in young participants with active exploration. Chapter Six summarises the findings of the research undertaken in my doctoral studies and discusses their implications for understanding the sensory mechanisms underlying tactile perception. Overall, the findings presented in this thesis suggest that the progressive loss of mechanoafferent units contribute to the decline in tactile spatial acuity as predicted by a population model of the afferent response, and provide new evidence on the complex effects of frictional changes and the role of skin biomechanics on the detection of a microdot.

Lung cancer is the second most diagnosed cancer, after breast cancer, worldwide. However, it is still the leading cause of cancer-specific mortality globally, contributing to 18% of all cancer-related deaths. Non-small cell lung cancer (NSCLC) makes up 85% of lung cancers and dependent on the stage, is amenable to a wide of treatments from surgery to systemic therapy. Immune responses within the tumour microenvironment have increasingly been implicated as determining factors in tumour progression and aggressiveness, and the focus has predominated on T-cell biology. The immune response is a complex interplay between the primary tumour and microenvironment, T and B cells. The role of the B cell in tumour survival is unclear but clearly has a function as tumour infiltration is commonly reported. Through deep phenotyping and multispectral tissue imaging techniques, we identified key differences in the effector and suppressive B cell composition between the tumour and peripheral blood compartments. IL10 positive suppressive B regulatory phenotypes were significantly more abundant in the circulation of patients who recurred post-operatively. Using a broad spectrum immunome array, we employed machine learning techniques and identified an auto-antibody signature in the serum of NSCLC patients that was highly predictive for post-operative recurrence in two independent cohorts. In addition to the techniques described above, we utilised functional ex vivo B cell assays to interrogate the response to checkpoint blockade in advanced disease patients and how this relates to B cell dynamics. Our findings demonstrated that lack of a suppressive B cell "brake" predisposed patients to high grade immune related adverse events post-treatment. Moreover, the B cells from toxicity patients were not only functionally defective in their ability to produce IL10 but also displayed a pan cytokine failure affecting pro-inflammatory cytokines thus suggesting B cell exhaustion in these patients. These findings significantly enhanced our understanding of the aetiopathogenesis of auto-immune toxicity secondary to checkpoint blockade with anti PD-1/PDL-1. In summary, this study aimed to explore the role of B cell biology in NSCLC by employing deep phenotyping and functional assay techniques at the blood and tissue level in both early and advanced stage disease. Our findings are likely to be informative in biomarker development for predicting response to treatment, post-operative relapse and for therapeutic adjuvant polyepitopic vaccine strategies in high-risk patients.

In Raman spectroscopy, data analysis occupies a large amount of time and effort; thus, it is paramount to have the proper tools to extract the most meaning from the Raman analysis. This thesis explores improved ways to analyse Raman data mostly by using machine learning techniques available in Python. The substrate used throughout this thesis has been patterned through an electrohydrodynamic process that patterns micrometric pillars onto the substrate, which, after being gold coated, can generate surface-enhanced Raman scattering. An initial theoretical background was laid for the electrohydrodynamic process and additional observations regarding the fluid mechanics. Furthermore, when the structures are fabricated, and Raman measurements are taken, we show that it is possible to create an effective convolutional neural networks that systematically evaluate these patterns' surface morphology and extracts features responsible for the surface-enhanced Raman scattering phenomenon. Being able to predict 90% of the time from optical microscope images and 99% of the time with atomic force microscopy images Additionally, a thorough machine learning analysis of the Raman literature was done. The best machine learning algorithms were put together into a script combined with a graphical user Interface that can run multiple commands such as principal component analysis and self-organizing maps, all in a centralised way. This way, we managed to consistently extract information from Raman and surface-enhanced Raman scattering spectra to open possibilities for precise peak analysis methods using a multi-Lorentzian fit algorithm.

Chronic liver diseases (CLDs) are characterised by inflammation and fibrosis which are driven by aberrant leukocyte recruitment. This process is mediated by specialised discontinuous endothelia, known as hepatic sinusoidal endothelial cells (HSEC), which act as the liver gatekeepers. During chronic inflammation, HSEC undergo substantial phenotypic changes, including upregulation of adhesion molecules, further driving leukocyte recruitment and exacerbating inflammation. The composition of the hepatic immune microenvironment determines the outcome of liver injury and understanding the mechanisms which regulate this process is critical to identify novel therapeutic targets. Mannose receptor (MR) and plasmalemma vesicle-associated protein (PLVAP) have previously been implicated in immune cell trafficking but their contribution to hepatic leukocyte recruitment remains undefined. This study aimed to characterise the expression of MR and PLVAP in normal and diseased human liver, understand their regulation in the hepatic sinusoids, and investigate their potential role in hepatic leukocyte recruitment. Immunohistochemistry and gene expression studies demonstrated that MR and PLVAP are differentially expressed in liver endothelium, with MR being downregulated and PLVAP being upregulated, in CLD and hepatocellular carcinoma. MR was resistant to regulation in primary human HSEC and did not seem to be involved in lymphocyte recruitment. In contrast, PLVAP was upregulated by several soluble factors, most notably by the senescence-associated secretory phenotype (SASP). The SASP, comprised of several cytokines and chemokines, is known to facilitate senescence surveillance by stimulating leukocyte recruitment. Additional experiments using patient tissue samples uncovered a previously unreported link between hepatic senescence, immune cell infiltration, and PLVAP expression in CLD. Furthermore, in vitro flow adhesion assays showed that PLVAP plays a selective functional role in monocyte paracellular transmigration, whilst having no impact on lymphocyte recruitment. These findings suggest that senescent cell-endothelial crosstalk drives PLVAP expression and shapes the immune landscape in chronic liver inflammation.

The global growth of solar power generation is rapid, yet the complex nature of cloud movement introduces significant uncertainty to short-term solar irradiance, posing challenges for intelligent power systems. Accurate short-term solar irradiance and photovoltaic power generation predictions under cloudy skies are critical for sub-hourly electricity markets. Ground-based image (GSI) analysis using convolutional neural network (CNN) algorithms has emerged as a promising method due to advancements in machine vision models based on deep learning networks. In this work, a novel deep network, "ViT-E," based on an attention mechanism Transformer architecture for short-term solar irradiance forecasting has been proposed. This innovative model enables cross-modality data parsing by establishing mapping relationships within GSI and between GSI, meteorological data, historical irradiation, clear sky irradiation, and solar angles. The feasibility of the ViT-E network was assessed the Folsom dataset from California, USA. Quantitative analysis showed that the ViT-E network achieved RMSE values of 81.45 W/m2 , 98.68 W/m2 , and 104.91 W/m2 for 2, 6, and 10-minute forecasts, respectively, outperforming the persistence model by 4.87%, 16.06%, and 19.09% and displaying performance comparable to CNN-based models. Qualitative analysis revealed that the ViT-E network successfully predicted 20.21%, 33.26%, and 36.87% of solar slope events at 2, 6, and 10 minutes in advance, respectively, significantly surpassing the persistence model and currently prevalent CNN-based model by 9.43%, 3.91%, and -0.55% for 2, 6, and 10-minute forecasts, respectively. Transfer learning experiments were conducted to test the ViT-E model's generalisation under different climatic conditions and its performance on smaller datasets. We discovered that the weights learned from the three-year Folsom dataset in the United States could be transferred to a half-year local dataset in Nottingham, UK. Training with a dataset one-fifth the size of the original dataset achieved baseline accuracy standards and reduced training time by 80.2%. Additionally, using a dataset equivalent to only 4.5% of the original size yielded a model with less than 2% accuracy below the baseline. These findings validated the generalisation and robustness of the model's trained weights. Finally, the ViT-E model architecture and hyperparameters were optimised and searched. Our investigation revealed that directly applying migrated deep vision models leads to redundancy in solar forecasting. We identified the best hyperparameters for ViT-E through manual hyperparameter space exploration. As a result, the model's computational efficiency improved by 60%, and prediction performance increased by 2.7%.

Camera based methods for optical coordinate metrology are growing in popularity due to their non-contact probing technique, fast data acquisition time, high point density and high surface coverage. However, these optical approaches are often highly user dependent, have high dependence on accurate system characterisation, and can be slow in processing the raw data acquired during measurement. Machine learning approaches have the potential to remedy the shortcomings of such optical coordinate measurement systems. The aim of this thesis is to remove dependence on the user entirely by enabling full automation and optimisation of optical coordinate measurements for the first time. A novel software pipeline is proposed, built, and evaluated which will enable automated and optimised measurements to be conducted. No such automated and optimised system for performing optical coordinate measurements currently exists. The pipeline can be roughly summarised as follows: intelligent characterisation → view planning → object pose estimation → automated data acquisition → optimised reconstruction. Several novel methods were developed in order to enable the embodiment of this pipeline. Chapter 4 presents an intelligent camera characterisation (the process of determining a mathematical model of the optical system) is performed using a hybrid approach wherein an EfficientNet convolutional neural network provides sub-pixel corrections to feature locations provided by the popular OpenCV library. The proposed characterisation scheme is shown to robustly refine the characterisation result as quantified by a 50 % reduction in the mean residual magnitude. The camera characterisation is performed before measurements are performed and the results are fed as an input to the pipeline. Chapter 5 presents a novel genetic optimisation approach is presented to create an imaging strategy, ie. the positions from which data should be captured relative to part's specific geometry. This approach exploits the computer aided design (CAD) data of a given part, ensuring any measurement is optimal given a specific target geometry. This view planning approach is shown to give reconstructions with closer agreement to tactile coordinate measurement machine (CMM) results from 18 images compared to unoptimised measurements using 60 images. This view planning algorithm assumes the part is perfectly placed in the centre of the measurement volume so is first adjusted for an arbitrary placement of the part before being used for data acquistion. Chapter 6 presents a generative model for the creation of surface texture data is presented, allowing the generation of synthetic butt realistic datasets for the training of statistical models. The surface texture generated by the proposed model is shown to be quantitatively representative of real focus variation microscope measurements. The model developed in this chapter is used to produce large synthetic but realistic datasets for the training of further statistical models. Chapter 7 presents an autonomous background removal approach is proposed which removes superfluous data from images captured during a measurement. Using images processed by this algorithm to reconstruct a 3D measurement of an object is shown to be effective in reducing data processing times and improving measurement results. Use the proposed background removal on images before reconstruction are shown to benefit from up to a 41 % reduction in data processing times, a reduction in superfluous background points of up to 98 %, an increase in point density on the object surface of up to 10 %, and an improved agreement with CMM as measured by both a reduction in outliers and reduction in the standard deviation of point to mesh distances of up to 51 microns. The background removal algorithm is used to both improve the final reconstruction and within stereo pose estimation. Finally, in Chapter 8, two methods (one monocular and one stereo) for establishing the initial pose of the part to be measured relative to the measurement volume are presented. This is an important step to enabling automation as it allows the user to place the object at an arbitrary location in the measurement volume and for the pipeline to adjust the imaging strategy to account for this placement, enabling the optimised view plan to be carried out without the need for special part fixturing. It is shown that the monocular method can locate a part to within an average of 13 mm and the stereo method can locate apart to within an average of 0.44 mm as evaluated on 240 test images. Pose estimation is used to provide a correction to the view plan for an arbitrary part placement without the need for specialised fixturing or fiducial marking. This pipeline enables an inexperienced user to place a part anywhere in the measurement volume of a system and, from the part's associated CAD data, the system will perform an optimal measurement without the need for any user input. Each new method which was developed as part of this pipeline has been validated against real experimental data from current measurement systems and shown to be effective. In future work given in Section 9.1, a possible hardware integration of the methods developed in this thesis is presented. Although the creation of this hardware is beyond the scope of this thesis.

This thesis explores whether an enterprise education competition (EEC), promoted as best practice vehicle of entrepreneurial education (EE), may have gendered outcomes in terms of differential impacts upon women participants, given mixed findings relating to the efficacy of EE for women which is considered to relate to the acknowledged masculine construction of entrepreneurship. Thus, this thesis investigates the influence of an EEC on the entrepreneurial self-efficacy (ESE), entrepreneurial intentions (EI) and perceived gender barriers to entrepreneurship of STEMM women early career researchers (ECRs). Of novelty, the theoretical framework draws upon Social Cognitive Career Theory (SCCT), from the career literature, which captures the socio-economic influences of perceived gender barriers upon entrepreneurial career intentions, while also exploring the predispositions of men and women EEC participants. Underpinned by a critical realist methodology, a quantitative study of 120 pre- and post-surveys of men and women participants, followed by 45 semi-structured interviews of women participants, were undertaken. The quantitative findings suggest that EEC participation removed the gender gap in: (a) perceived stereotype threat, (b) perceived networking difficulty and (c) perceived ability related to entrepreneurial finance and cost estimation. However, following EEC participation, women participants continued to perceive high barriers in: (a) sex discrimination, (b) childcare-work conflict and (c) a lack of role models and mentors when compared to their male counterparts. In addition, the EEC deterred self-confidence in ESE of women participants who perceived a high barrier in stereotype threat. Despite this, evidence from the qualitative study suggests that the EEC programme still reproduced unintended gendered outcomes for some women participants, particularly in: (a) perpetuating the stereotypical masculine stereotypes of an entrepreneur, (b) reinforcing perceived conflict between childcare and work-life of a woman entrepreneur as well as (c) highlighting a negative image of STEMM women entrepreneurs.

There is a growing interest in assessing the impact of drivers' actions and behaviours on road safety due to the numerous road fatalities and costs attributed to them. For Heavy Goods Vehicle (HGV) drivers, assessing the road safety risks of their behaviours is a subject of interest for researchers, governments and transport companies, as nations rely on HGVs for the delivery of goods and services. However, HGV driving is a complex, dynamic, uncertain and multifaceted task, mostly influenced by individual traits and external contextual factors. Advanced computational and artificial intelligence (AI) methods have provided promising solutions to automatically characterise the manner by which drivers operate vehicle controls and assess their impact on road safety. However, several challenges and limitations are faced by the current intelligence-supported driving risk assessment approaches proposed by researchers, such as: (1) the lack of comprehensive driving risk datasets; (2) information about the impact of inevitable contextual factors on HGV drivers' responses is not considered, such as drivers' physical and mental states, weather conditions, traffic conditions, road geometry, road types, and work schedules; (3) ambiguity in the definition of driving behaviours is not considered; and (4) imprecision of AI models, and variability in experts' subjective views are not considered. To overcome the aforementioned challenges and limitations, this multidisciplinary research aims at exploring multiple sources of data including information about the impact of contextual factors captured from crucial stakeholders in the HGV sector to develop a reliable context-aware driving risk assessment framework. To achieve this aim, AI methods are explored to accurately detect drivers' driving styles, affective states and driving postures using telematics data, facial images, and driver posture images respectively. Subsequently, due to the lack of comprehensive driving risk datasets, fuzzy expert systems (FESs) are explored to fuse detected driving behaviours and perceived external factors using knowledge from domain experts. The key findings of this research are: (1) recurrent neural networks are effective in capturing the temporal dynamics and differences between the different types of driver distraction postures and affective states; (2) there is a trade-off between efficiency and privacy in processing facial images using AI approaches; (3) the fusion of driver behaviours and external factors using FESs produces realistic, reliable and fair driving risk assessments; and (4) a hierarchical representation of a decision-making process simplifies reasoning compared to flat representations.

The Northern Cape Province of South Africa is host to some of the largest deposits of iron and manganese metal in the world, contained in two separate districts, namely the Kalahari (KMF) and Postmasburg Manganese fields (PMF). The current thesis attempts to shed light on the chronicle of the complex epigenetic history of the two ore fields by focusing on the previously underexamined recurring theme of gangue alkali-rich assemblages, contained in and flanking a series of Fe and Mn orebodies. The principal outcomes of this study not only (a) set constrains on the characteristics, origin, sources, timing and relationship of alkali mineralization between the KMF and PMF, but also (b) challenge current ore genetic models by exploring the role of regional "alkali flooding" as a shared mechanism in the formation or at least hydrothermal upgrade of the deposits, (c) are beneficial for future exploration for new orebodies in the study and other prospective areas and (d) improve our knowledge of fluid circulation in sedimentary basins.

This research explores the application of manual facial comparison methods and offline (semi-) automated face recognition algorithms to human identification of recently deceased individuals, using 3D (N=3) and 2D (N=6) data. Current methods are generally designed for and applied to the living, are not commonly used in combination or hierarchical order, and are mostly tested under controlled, lab-based conditions. The aforementioned makes their applicability to real-life settings questionable, and research on their application to the deceased is scarce, although desperately needed. Further issues arise from the lack of uniformity and standardisation in methodological approaches, as well as feature descriptions and terminology. This study investigates the applicability of combined, manual face comparison methods in hierarchical order (preliminary feature-based analysis, facial superimposition, detailed morphological comparison) and two (semi-) automated face recognition algorithms (MATLAB® and Picasa) to the recently deceased. Pilot and ancillary studies explore pretend-dead faces as a data source and evaluate geometry vs. texture in 3D face models. 3D data was obtained using the Artec handheld laser scanner. Key findings indicate that human face matching ability is superior and more resilient to PM facial changes, non-standardised AM data, and limited data availability, compared to the automated methods tested. Results further suggest that a hierarchical approach to manual comparison is highly beneficial. MATLAB®'s algorithm is unreliable even as a filtering tool and Picasa struggled to detect PM faces in images; an issue not encountered with pretend-dead data. Both automated approaches utilised here are not suitable for application on similar datasets or in casework settings, unlike the manual approach - the latter requiring further validation. Limitations arise primarily from the small sample size and non-quantifiable approaches to manual facial superimposition. In forensic casework and disaster victim identification scenarios, comparative AM data for primary and even secondary methods of identification are often lacking. However, facial photographs are almost always attainable and should be considered an important resource for post-mortem identification. The interdisciplinary nature of this field requires collaborative efforts to address remaining challenges in the future.

Research on human populations and dependency on coral reefs is relatively sparse, and often uses recycled statistics that have not been updated for many years. In light of climate change on coral reefs, its vital that continual assessments not only on coral reef ecosystems, but additionally the communities that rely on them is maintained in order to inform climate resilience. I aimed to 1) update statistics and create a long-term dataset of human populations near coral reefs, and 2) develop a human dependency framework that is reproducible and adaptable to newly available data. Using openly accessible data, of LandScan and global coral reef distribution map, I found that nearly 1 billon people live within 100 km of coral reefs in 2020. I developed a conceptual human dependency framework, which encompassed four pre-defined dependency categories, of fisheries, tourism, coastal protection and nutrition. Using an indicator approach, openly accessible data was collected for each category. The Human dependency on coral reef index (HDCRI) was developed and calculated, and were complemented by hybrid learning techniques. Human dependency "profiles" were created, and presented, how countries were dependent on coral reef ecosystems and what indicators were driving the dependency. The conceptual framework, aimed to shift thinking of human dependency on a linear scale from low to high dependency, towards a more holistic view on human dependency. The human dependency framework and population methods, were designed to be reproducible and adaptable to different scales of data (e.g. regional and national levels), and updated with improved datasets. The outputs of these studies are aimed to improve coral science by facilitating human aspects to research, additionally, to create more informed decision making to policymakers in distributing funds and resources. Finally, to facilitate a novel tool of insurance as a form of climate resilience, for coral reefs and the humans that depend on them.

Low power mobile computing systems such as smartphones and wearables have become an integral part of our daily lives and are used in various ways to enhance our daily lives. Majority of modern mobile computing systems are powered by multi-processor System-on-a-Chip (MPSoC), where multiple processing elements are utilized on a single chip. Given the fact that these devices are battery operated most of the times, thus, have limited power supply and the key challenges include catering for performance while reducing the power consumption. Moreover, the reliability in terms of lifespan of these devices are also affected by the peak thermal behaviour on the device, which retrospectively also make such devices vulnerable to temperature side-channel attack. This thesis is concerned with performing Dynamic Voltage and Frequency Scaling (DVFS) on different processing elements such as CPU & GPU, and memory unit such as RAM to address the aforementioned challenges. Firstly, we design a Computer Vision based machine learning technique to classify applications automatically into different categories of workload such that DVFS could be performed on the CPU to reduce the power consumption of the device while executing the application. Secondly, we develop a reinforcement learning based agent to perform DVFS on CPU and GPU while considering the user's interaction with such devices to optimize power consumption and thermal behaviour. Next, we develop a heuristic based automated agent to perform DVFS on CPU, GPU and RAM to optimize the same while executing an application. Finally, we explored the affect of DVFS on CPUs leading to vulnerabilities against temperature side-channel attack and hence, we also designed a methodology to secure against such attack while improving the reliability in terms of lifespan of such devices.

Epigenetic mechanisms govern gene regulation and respond to both genetic and environmental influences. Epigenetic marks, including DNA methylation (DNAm) are dynamic across an individual's life and may be influenced by several environmental and genetic factors. In this thesis, I evaluated several factors which influence the human epigenome using a bioinformatics approach. First, I explored the idea that DNAm changes may underlie the link between air pollution and poor health. I identified no significant link between background air pollution levels and DNAm, however, was able to identify a link between exposure to traffic related air pollution (TRAP) and DNAm. I identified 531 significant CpG sites related to TRAP which were enriched at regulatory sites and novel and previously reported transcription factor (TF) motifs and genes. Secondly, I explored the role of DNA methylation in autosomal sex differences. Specifically, I identified and validated 396 CpG sites and 266 differentially methylated regions associated with sex. We found the majority of these to be female-biased CpGs (74\%) and were enriched in CpG islands and located preferentially at 5/3'UTRs and enhancers. TF motif enrichment revealed enrichment of TF's related to critical developmental processes and sex determination such as SRY and ESR1. Finally, I report a catalogue of loci across the human epigenome displaying either variable or stable inter- individual DNA methylation. I demonstrate that the majority of the VMPs are not controlled by age, sex or smoking status. I also report highly variable and stable CpG sites enriched at methylation sensitive TFs, highlighting that there is potentially a strong relationship between DNAm and TF binding, which will be important for gene regulation. In addition, I found that the VMPs were under higher genetic control than the SMPs and that this is in part directed by interindividual differences in 3D chromatin organisation. In summary, the results presented in this thesis give insights into the role that external factors play in influencing the human epigenome.

Cancer is a major burden upon society and a leading cause of death world-wide. In trying to understand the development and progression of cancer better, we want to identify the events in a tumour sample that are absent in normal tissue. Our understanding of tumour development and progression could be greatly aided through transcriptomic analysis of matched tumour and normal tissues. However, sequencing of such tissues have been hampered by issues with RNA quality (i.e. samples are often degraded especially samples that have been stored under non-optimal conditions and/or fixed). Such datasets, making use of matched samples are therefore rare. However, as the study of small RNAs became more and more popular, some matched sample analysis exists to study microRNA expression. We hypothesise that degraded RNA from clinical samples may be present in small RNA libraries, and could potentially be used to study the expression level of protein coding transcripts. Small RNA-seq was analysed to see if gene expression changes could be predicted from the degraded RNA present in the small RNA pool. First, I processed small RNA datasets with modified mapping and annotation pipelines to analyse total RNA, and compare their expression values with the measured levels from standard total RNA libraries. Different transformation methods were considered and for all of them I found a strong association between expression levels from total RNA libraries and RNA from small RNA libraries. Through analysis of the degraded RNA from patient tumour samples, I was able to identify genes differentially expressed including FASN, CAMKII, CREB3L4, AR, PAK1, FUT8, FKBP5. The gene expression changes quantified from degraded RNA with alterations were compared with total non-degraded RNA. Some of the genes found to be regulated in the total RNA were also found to be regulated in the degraded RNA including TIMP3, PGR, AREG, AMZ1. The data demonstrated a strong association between gene expression analysis performed on the degraded (small) RNA and total RNA.

Globally, populations are ageing, and birth rates are falling. This presents a variety of issues, including food shortages and a lack of working-age humans to sustain the older generations. This presents a global requirement for research into human and animal fertility: animal research to help with global food security allowing a larger population to be fed more effectively, and human research to help tackle the declining fertility rates to ensure that the population continues to grow at a level that can support the larger number of elderly people. The area of fertility research that needs to be investigated most is that of sperm development. There is traditionally a lot more female-factor research, leaving a knowledge gap for male-factor research. Sperm development in mice has been shown to be directly linked to successful fertilisation outcomes: mouse sperm with perturbed development do not produce viable embryos. In some cases, this has been shown to be a sex-I inked effect, whereby female mouse embryos are produced, but male mouse embryos are not. This study also shows that autosomal genes are critical for proper mouse sperm development, and so differential gene expression analysis methods are used to determine which genes have a greater impact: autosomal, or sex-linked, and to determine which specific genes may be the most implicated in poor mouse sperm development. The list of candidate genes found in this study to potentially be involved in poor mouse sperm development provides possible targets for gene therapy to improve fertilisation outcomes, and orthologues of these genes can be investigated in agricultural species such as pigs and cows, as well as humans.

The scarcity of labelled data is a serious problem since deep models generally require a large amount of training data to achieve desired performance. Data augmentation is widely adopted to enhance the diversity of original datasets and further improve the performance of deep learning models. Learning-based methods, compared to traditional techniques, are specialized in feature extraction, which enhances the effectiveness of data augmentation. Generative adversarial networks (GANs), one of the learning-based generative models, have made remarkable advances in data synthesis. However, GANs still face many challenges in generating high-quality augmented images from small datasets because learning-based generative methods are difficult to create reliable outcomes without sufficient training data. This difficulty deteriorates the data augmentation applications using learning-based methods. In this thesis, to tackle the problem of labelled data scarcity and the training difficulty of augmenting image data from small datasets, three novel GAN models suitable for training with a small number of training samples have been proposed based on three different mapping relationships between the input and output images, including one-to-many mapping, one-to-one mapping, and many-to-many mapping. The proposed GANs employ limited training data, such as a small number of images and limited conditional features, and the synthetic images generated by the proposed GANs are expected to generate images of not only high generative quality but also desirable data diversity. To evaluate the effectiveness of the augmented images generated by the proposed models, inception distances and human perception methods are adopted. Additionally, different image classification tasks were carried out and accuracies from using the original datasets and the augmented datasets were compared. Experimental results illustrate the image classification performance based on convolutional neural networks, i.e., AlexNet, GoogLeNet, ResNet and VGGNet, is comprehensively enhanced, and the scale of improvement is significant when a small number of training samples are involved.

Cell therapies hold the potential to halt or reverse several pathologies, including Parkinson's Disease. Despite showing promise, cell death post transplantation remains a limitation and recent findings have suggested that a paucity of glucose is playing a crucial role in graft death. To improve graft survival, this work aimed to formulate glucose releasing materials, which can sustain cell metabolism during hypoxic period, when cells are facing a condition of oxygen-glucose deprivation (OGD). To assay the glucose released from the various formulations tested, an enzymatic assay based on glucose oxidase/horseradish peroxidase tandem reaction was optimised for glucose detection in the 0.2-2 mg/mL range, relevant to cell culture techniques. By encapsulating glucose into a variety of photocrosslinkable acrylate-containing monomers, it was possible to produce macroscale version (1.5 x 2.0 mm) of the glucose releasing materials. These materials allowed for an easy screening of several formulations, which were used to understand the principles underlaying glucose release and how to obtain sustained glucose delivery. Although it was possible to formulate injectable version of these materials, the glucose release rate from the microparticles was mostly brief and inadequate to the time course required. Nonetheless, one of the formulations suggested that sustained glucose release could be obtained also from microparticles. To mimic the prolonged graft ischemia, and assess glucose impact on hypoxic neurons, iPSCs derived neurons (dopaminergic, medium spiny and cortical), foetal ventral mesencephalon tissue and SH-SY5Y cells were cultured up to 96 hours either at 21% or 0.1% pO2 at different glucose concentrations. iv This work shows that: 1) glucose paucity, rather than oxygen deprivation, negatively affects neurons viability; 2) glucose can be released in a sustained manner (7-16 days) and, although showing a brief release, 3) in vitro, the microparticles were able to rescue cell viability from OGD-induced death.

Habitat loss, degradation, and fragmentation present significant threats to biodiversity globally. Amphibians are among the most threatened vertebrates, yet the available conservation literature covering the effects of habitat fragmentation on amphibians, especially using genetic methods, is not comprehensive. Within this thesis, a combination of field data, mitochondrial DNA barcoding and next generation sequencing were integrated to evaluate ecological and evolutionary mechanisms that influence current patterns of amphibian diversity across a fragmented landscape in Madagascar's Central Highlands. Amphibian diversity and community composition differed in forest fragments, both influenced by fragment size and isolation. Species and phylogenetic diversity were diminished and communities in fragments were subsets of those in more continuous forest. Diversity was lost across the phylogeny, as opposed to losing entire clades of closely related species, with the naturally rare and terrestrial species most commonly lost. Our results suggest that local forest fragmentation is recent, with evidence of extinction debt, but that the Central Highlands has long been a matrix of savannah grasslands and forest. Consequently, this recent habitat loss timeline may not have been sufficient to negatively influence genetic diversity, and I conclude that the absence of some species in the smaller fragments is likely due to ecological factors as opposed to genetic erosion. My results add knowledge of the cryptic diversity remaining in forest fragments in this region and demonstrate the importance of conserving and re-connecting the remnant forest fragments. This thesis expands the understanding of the effects of habitat fragmentation on amphibian diversity and provides the first use of phylodiversity and population genetics to examine the impacts of habitat fragmentation on amphibians in Madagascar. The results of this thesis can directly be used as a baseline for future conservation and management of the amphibians studied, and our results can guide reforestation efforts in the region.

Accumulating evidence suggests an emerging association between perturbed haematopoiesis, development of leukaemia and cardiovascular disease in the context of a high-fat western diet. To explore this subject, I investigated the role of atherosclerosis prone low-density lipid receptor (Ldlr) in normal and leukaemic haematopoiesis and the impact of a high-fat diet (HFD) or normal chow diet (ND) in this setting. In steady-state, under normal dietary conditions, mice engineered to be deficient in the LDL receptor (Ldlr-/-) had increased numbers of haematopoietic stem cells (HSCs), which was associated with increased cell cycling and an increase in inflammatory cytokines and chemokines. In Ldlr-/- mice bone marrow differentiation, as assessed by the CFC assay, was decreased while paradoxically white blood cells were increased which mapped to CD4+ T cell and monocyte increases in the peripheral blood. To induce hypercholesteremia and atherosclerosis, Ldlr-/- mice were fed a HFD and the attendant impact on haematopoiesis was evaluated. A significant increase in HSCs and associated early progenitor compartments (HSPCs) was noted in Ldlr-/- mice fed HFD alongside an increase in committed progenitor cells of both the myeloid and lymphoid lineage. As expected, inflammatory immune cell subsets were increased together with increases in platelets and alterations in regulatory immune cells in Ldlr-/- mice fed HFD. HSCs from Ldlr-/- mice fed HFD performed poorly in functional analysis, as judged by competitive transplantation, displaying significant multi-lineage differentiation defects. Underpinning these defects, RNA-seq analysis revealed altered apoptosis, inflammation, lipid metabolism pathways, RNA biology, and AML enriched gene pathways in HSCs from Ldlr-/- mice fed a HFD. These molecular pathways mapped not only to haematological disease, like AML, and cardiovascular disease, but also nephrotoxicity and hepatoxicity, highlighting the widespread impact of perturbed haematopoiesis induced by HFD and atherosclerosis. Unexpectedly, we found that MLL-AF9 transformed HSPCs from Ldlr-/- mice fed a HFD developed AML later than their ND counterparts, but this was likely reflected by a delayed migration of leukaemic blast cells from BM to PB. This argument was supported by altered adhesion protein expression in human MLL-AF9 AML cell lines exposed to atherogenic lipoproteins in vitro. Decreased markers of immune recognition were also observed in human MLL-AF9 AML cell lines exposed to atherogenic lipoproteins in vitro. The data provided in this thesis provide mechanistic vi insights into how HFD epigenetically disrupts HSC function and haematopoiesis in the setting of atherosclerosis, and it provides a starting point to further explore relationship between HFD, atherosclerosis and how perturbed haematopoiesis can lead to AML.

Bdellovibrio bacteriovorus is an endobiotic bacterial predator that encodes multiple enzymes related to known peptidoglycan (PG) modifiers, which are necessary for predation to occur. Additionally, a plethora of cryptic proteins are expressed and upregulated during predation. It has been shown that L,D-transpeptidases (LDTs) are required for bdelloplast formation and stabilisation. These LDTs cluster in six groups in B. bacteriovorus HD100, nevertheless, no further LDT information has been previously described. Here we have determined and analysed the structures of LDTs Bd0566, Bd1075, Bd3334 and Bd3741, showing that even though they share a structurally conserved catalytic domain, they have acquired accessory domains and motifs characteristic to each LDT group. Moreover, we have integrated these novel structures with the previously solved structures of Bd0553, Bd0886 and Bd1402, generating a panel of seven LDTs representing the six groups of B. bacteriovorus HD100, allowing us to propose possible roles and the structural determinants defining the transferase/crosslinking or hydrolytic activities of each group. Further, we report the structure of two predatory NlpC/P60 amidohydrolases/peptidases, Bd0601 and Bd1177. We have determined that Bd0601 is a DUF1460 amidohydrolase, wherein the PG stem peptide binding groove is blocked by a tyrosine residue, suggesting the need of a conformational rearrangement for substrate binding. On the other hand, we have determined that Bd1177 is a bilobular NlpC/P60, wherein the active site is located in the interface between the catalytic N-terminal domain and a novel C-terminal regulatory domain, suggesting that Bd1177 fluctuates to an open conformation upon substrate binding. Finally, we have determined the crystallographic structure of the cryptic secreted protein Bd0675, depicting a novel fold. The solved structure depicts a prominent central acidic and aromatic groove delimited by a conserved disulphide-stabilised loop, suggesting that this protein has ligand binding functions. Based on the solved structure we have tested different candidate ligands; however further experiments are required to propose a function for Bd0675. In addition, due to its novelty, Bd0675 was used as target for multiple structure predicting algorithms in CASP14.

This thesis describes how modifying PDMS can influence the polymer's sorbent properties towards organic compounds in aqueous matrices. Where each new sorbent's extraction performance was measured through TD-GC-MS and characterized through various analytical techniques. Chapter 1: Gives a brief introduction to the analytical techniques used within the thesis whilst outlining the current state-of-the-art PDMS modification techniques reported in recent literature. Chapter 2: Provides the detailed design of experiment and the corresponding materials used for each results chapter. Chapter 3: This chapter shows the results and discussions section for the encapsulation of amines with PDMS. In which each PDMS-based material demonstrated variable sorptive extraction properties towards organic compounds in aqueous solutions compared to PDMS. However, each material presented thermal instability and extraction performance dropped over time. Chapter 4: Highlights how bonding amines to PDMS improved the thermal stability of the sorbent material. However, the TD-GC-MS results showed little variance in extraction performance compared to that of the currently used PDMS, unlike that of the encapsulated method. Chapter 5: This section looked at how a novel preparation technique aimed to incorporate established commercially available sorbents within the PDMS. These bi-phasic sorbents were tested for organic compound extraction via TD-GC-MS. The results showed that the PDMS-Tenax GR outperformed all other sorbents for most of the tested organic compounds. Whilst demonstrating both chemical and thermal robustness. The main issue around this section was how to scale-up the material preparation step. Chapter 6: This chapter provides an overview of the results obtained during the thesis while highlighting the areas of improvement. Future works state the direction this research should move based on the result obtained.

Introduction: Response to neoadjuvant chemoradiotherapy (NCRT) for locally advanced rectal cancer (LARC) varies considerably. There is a need to unravel the elusive biological mechanisms behind treatment resistance in LARC and discover radiosensitising treatments using the latest experiment models. Methods: Next generation sequencing (NGS) was performed on archival specimens from 23 LARC patients (retrospective cohort) to identify differentially expressed genes associated with NCRT response. Six patient derived organoid (PDO) models were derived (prospective cohort) from colorectal cancer (CRC) patients; genetically and immunohistochemically characterised. In vitro viability assays were conducted to determine PDO response to radiotherapy. NGS was performed on PDOs pre- and post-irradiation. Chemoradiotherapy viability assays using targeted pathway inhibitors were performed using HCT116 CRC cell line and PDOs. AKT phosphorylation following chemoradiotherapy was assessed using western blots. Results: The first 6 out of 16 CRC PDO lines successfully derived in the laboratory were characterised through genomics and immunohistochemistry. Several genes and biological pathways of interest in radiotherapy response (sensitivity or resistance) were identified on differential expression analyses and Gene Set Enrichment Analysis of the retrospective FFPE sample and prospective organoid sample transcriptomes. The PI3K/AKT/mTOR pathway upregulation was associated with radiotherapy resistance in retrospective and prospective cohort sample transcriptomic analyses. Radiotherapy was associated with significantly increased AKT phosphorylation in HCT116. The use of PI3K and mTOR dual inhibitors apitolisib and dactolisib radiosensitised HCT116 and PDOs in vitro and led to inhibition of radiation induced AKT phosphorylation. These drugs radiosensitised radioresistant PDO lines and HCT116 with maximal inhibitory concentration levels within previously published ranges for humans. Dual inhibitors may also possess chemotherapy sensitising properties in the absence of radiotherapy. Conclusion: The PI3K/AKT/mTOR pathway upregulation is associated with NCRT resistance. The role of dual PI3K and mTOR inhibitors as radiosensitisers in LARC patients warrants further preclinical and clinical research.

The accretion of oceanic crust at mid-oceanic ridges (MOR) accounts for the most voluminous magmatism on Earth and occurs across a range of spreading-rates. Slow-spreading ridges represent 50% of the present-day global MOR-system and produce heterogeneous oceanic crust that deviates fundamentally from the conventional 6 - 7 km layer-cake Penrose-crust formed at fast-spreading ridges. Instead, the sparse and intermittent magmatism at slow-spreading ridges requires plate separation to be partly accommodated by faulting, producing ocean lithosphere with a discontinuous igneous crust of variable thickness, commonly disrupted by large 'detachment' faults that exhume the deep crust and shallow mantle directly onto the seafloor. Due to the difficulty in accessing oceanic crust in situ, our knowledge of the processes occurring at MORs is still relatively limited and geologists often turn to ophiolites as their analogues. There is however an increasing awareness of a preservation bias towards ocean lithosphere formed at atypical spreading ridges, such as marginal ocean basins near subduction zones, and that potentially key differences between many ophiolites and 'true' MOR ocean lithosphere (especially in their chemical composition) complicate direct comparisons. The Masirah Ophiolite, exposed over ~650 km2 on an isolated island off the southeast coast of the Sultanate of Oman, is near-unique in that it is believed to have formed at a 'true' MOR, unaffected by the influence of subduction. It can therefore provide valuable geological insights into crustal accretion process along modern MORs. Although its basic structure was mapped the 1990s, previous studies did not examine Masirah from the perspective of modern MOR processes in any detail. Previous work determined that Masirah formed at a slow-spreading ridge in the young Indian Ocean at ~150 Ma, followed by an episode of alkaline magmatism during intraplate rifting some 20 Ma later, before being emplaced onto the Arabian continental margin in the late Cretaceous. The work presented in this thesis updates the geochronological model for the evolution of Masirah, describes a previously unrecognised type of ocean lithospheric architecture ('Penrose on a diet') that accreted at the paleo-spreading ridge, and investigates what mantle processes might be responsible for this style of accretion. In a key finding, new radiometric dates show that the ophiolite formed at 135 - 130 Ma, and that the alkaline magmatic activity overlapped with crustal accretion as an episode of 'near-axis' magmatism. Trace element compositions of the two magmatic suites show a large degree of overlap and define a continuum, supporting a transitional event where magmas, derived by variable degrees of melting of a heterogeneous mantle source, were initially delivered to the ridge axis, whereupon subsequent melts, derived from continued lower-degree melting of a more enriched mantle component, were delivered near off-axis. The 'Penrose on a diet architecture' of the Masirah lithosphere is noteworthy for having a thin igneous crust (2 km) and a thin lower crust with respect to the upper crust (lower crust : upper crust = ~0.4 : 0.6). Despite these characteristics suggesting a low melt supply, field relations nevertheless indicate the crust formed by magmatic spreading. This contradicts current crustal accretion models, which predict the formation of 6 - 7 km thick crust for conditions of high melt supply and spreading dominated by detachment faulting when melt supply is low. The Masirah Ophiolite shows that the variability of permissible slow-spreading crustal geometries is greater than previously thought and that a basaltic seafloor should not automatically be interpreted to correlate with a thick magmatic crust. The mantle section exposed on Masirah is characterised by highly refractory peridotites with a mineral chemistry indicating high degrees of melting, despite the overlying igneous crust being thin. By considering the effect of mantle fertility on the overall melt production, a model is advanced where upwelling of an ancient domain of depleted mantle led to an overall reduced melt supply and a high proportion of melts from fertile mantle components relative to those from the depleted peridotites. Consequently, observations from Masirah suggest that besides rates of upwelling and seafloor spreading, the time-integrated melting history of the mantle underneath modern MORs exercises an important control on the degree of mantle melting, as well as the resulting lithospheric architecture and compositions of the erupted mid-ocean ridge basalts.

The patient is the expert of their medical journey and their experiences go largely unheard in clinical practice. Understanding the patient is important as bridging gaps in the medical domain enhances clinical knowledge, benefiting patient care in addition to improving quality of life. Valuable solutions to these problems lie at the intersection of Machine learning and sentiment analysis; through ontologies, semantic similarity, and clustering. In this thesis, I present challenges and solutions that explore patient quality of life pertaining to two inflammatory diseases: Uveitis and Inflammatory Bowel Disease, which are immune-mediated inflammatory diseases and often undifferentiated. This thesis explores how a patient's condition and inflammation influences their voice and quality of life via sentiment analysis, clustering, and semantic characterisations. Methods With guidance from domain experts and a foundation derived from clinical consensus documents, I created an application ontology, Ocular Immune-Mediated Inflammatory Diseases Ontology (OcIMIDo), which was enhanced with patient-preferred terms curated from online forum conversations, using a semi-automated statistical approach - with application of annotating term-frequency and sentiment analysis. Semantic similarity was explored using a preexisting embedding model derived from clinical letters to train other models consisting of patient-generated texts for systematic comparison of the clinician and patient voice. In a final experimental chapter, blood markers were clustered and analysed with their corresponding quantitative quality of life outcomes using patients in the UK Biobank with Inflammatory Bowel Disease. Results OcIMIDo is the first of its kind in ophthalmology and sentiment analysis revealed that first posts were more negative compared to replies. Systematic comparisons of embedding models revealed frequent misspellings from clinicians; use of abbreviations from patients; and patient priorities - models performed better when the clinical domain was extended with equivalent-sized, patient-generated data. Clusters unveiled insight into the presence of inflammatory stress and the relationship with happiness and the presence of a maternal smoking history with a Crohn's disease diagnosis. Summary Patient-preferred terms prove the patient voice provides meaningful text mining and fruitful sentiment analysis, revealing the role a forum plays on patients; semantic similarity highlighted potential novel disease associations and the patient lexicon; and clustering blood markers featured clusters presenting a relationship with sentiment. In summary, this deeper knowledge of quality of life biomarkers through the patient voice can benefit the clinical domain and patient outcomes as understanding the patient can improve the clinical-patient relationship and communication standards: all benefiting the diagnosis process, developing treatment plans, and shortening these intensive time hauls in clinical practice.

The Synthetic Aperture Radar (SAR) is a well-known imaging radar. Whilst its technology is proven, it is, however, subject to improvement as technology improves, just like any other technical system. With the use of multiple antenna, high-resolution wide swath (HEWS) imaging has advanced dramatically, a duty that MIMO radars are currently fulfilling. Parallel to this, co-located, linear MIMO arrays have proliferated; they are already utilised for automobile safety features, and their compact form factor allows them to be deployed on drones. These platforms are unique in that they frequently require contiguous azimuth swaths (for example, aerial field surveying or monitoring road conditions), but they also have the potential to perform high-resolution imaging due to the high operating frequency of commercial MIMO arrays (24 and 77 GHz) and the short stand-off distances. MIMO-SAR systems make use of newly developed SAR systems. A new branch of MIMO-SAR imaging has emerged, utilising high-frequency low-altitude UAV/drones. A novel model, high-resolution contiguous swath could potentially fill a gap in this sector (ARCS). The concept specifically employs Digital Beamforming (DBF) techniques, which are enabled in such systems to give contiguous azimuth imaging, as in stripmap SAR, but with a fine spatial resolution comparable to spotlight SAR. Strip-Spot SAR, which we have devised as a new concept, was initially examined as SAR and MIMO-SAR, after which the concept's key principles were analytically defined and empirically validated under laboratory conditions with calibrated targets and extended targets. It is hoped that it will be able to serve as a model for future research on contiguous swath imaging.

Titanium (Ti) and its alloys are attractive metals for biomedical use, where predominantly Ti-6Al-4V (Ti64) is used. This is largely due to their desirable mechanical and chemical properties, which include low density, excellent fatigue resistance, and high corrosion resistance. Additionally, Ti is renowned for its excellent biocompatibility and for being one of the only known metals to promote osseointegration, which is the growth of bone onto the surface of the metal. However, untreated Ti suffers from very poor wear resistance that results in the release of wear particles and ions into the surrounding tissue and systemic circulation. This is worsened with alloys like Ti64 due to the presence of vanadium, which is known to be highly toxic, so alternative alloys are being investigated, one of which is Ti-6Al-7Nb (Ti67), which has also been investigated within this study. Ceramic conversion treatment (CCT, C2T) is an effective approach used to resolve the poor tribological properties. This thermal oxidation process results in the formation of a hard, wear resistant surface rutile (TiO₂) oxide layer, with an underlying oxygen-hardened diffusion zone (ODZ). Additionally, CCT further improves the corrosion resistance of Ti, which is beneficial in biological applications due to the aggressive in vivo environment. However, it takes approximately 80-100 hours to produce a high-quality oxide layer of 2-5 µm on Ti64. The research within this thesis is focused on enhancing the CCT process through surface pre-deposition of catalytic films consisting of Au, Ag, Pd, or a combination of Ag and Pd to produce novel catalytic ceramic conversion treatment (CCCT, C3T). Detailed characterisation of the influence of these coatings on the CCT process demonstrate a significant increase in oxide layer thickness from ≈ 5 to ≈ 100 µm with the same treatment conditions, and a reduction in treatment time from 80 to 1 h to generate an equal layer thickness. C3T presented an improved oxide layer-substrate inter-facial bonding, aluminium-rich layers at the subsurface, and surface agglomeration. In-depth analysis of C3T Ag/Pd by TEM and EDX revealed a surface layer consisting of 2 sublayers: (1) a superficial layer containing agglomerates of palladium (and trace silver) oxide, and amorphous-like alumina fine grains within the grain boundaries. (2) an underlying structure dominated by nano-columnar grains of TiO₂Al₂O₃, with trace silver and palladium particles. Owing to the alumina produced at the subsurface and the underlying supporting nano grains, the C3T layers were of higher surface hardness and load bearing capacity, and therefore demonstrated a significant improvement in durability during reciprocating tribological testing. A potential application for the novel C3T was assessed by fixation pin drilling into a simulated bone material with untreated and treated pins. It revealed the resilience of the oxide layers even when produced on complex geometry objects, where the treatments both reduced insertion force and the generated temperatures during drilling.

Tuberous sclerosis complex (TSC) is caused by mutations in the genes TSC1 or TSC2 and is a rare genetic disorder. These mutations result in loss of function of the respective gene, thus leading mammalian target of rapamycin complex 1 (mTORC1) overactivity. TSC is a multi-organ disorder with known comorbidities such as glia tumours, epilepsy, autism spectrum disorder (ASD) or psychiatric disorders. Despite this, it is still unknown what molecular and developmental changes are occurring in the patient brain. Alzheimer's Disease (AD) is neurodegenerative disease and the most common form of dementia. Literature research suggests that similar molecular pathways are dysregulated in both AD and TSC and this project is aiming to identify commonalities in the developmental, autophagy and inflammation related markers. Human induced pluripotent cells (iPSCs) enable human astrocytes and neurons to be cultured in vitro, thereby circumventing developmental and genetic differences of an animal model compared to humans, and they allow the generation of neurons and astrocytes; cell types which are otherwise inaccessible from humans due to major ethical concerns unless post-mortem. By using two CRISPR-based techniques CRISPR-Cas9 and CRISPR-Cas13a systems, both knockout and knockdown models for TSC1 were generated. These cell lines underwent neuronal differentiation to identify changes in developmental and autophagy markers similar to AD and a database analysis of TSC patients identified further target genes/pathways which were common between both disorders. The TSC1 cell models were used to generate neurons to investigate the effect of TSC1 loss on neuronal development. While the knockout model creates a chronic loss of TSC1, the knockdown model enabled controlled reduction of TSC1 mRNA at different timepoints of development for brief periods. Expression of developmental, inflammation, autophagy and urokinase pathway marker genes were tested at different timepoints of the neuronal development. Astrocytes generated from Cas13 cells, where the TSC1 knockdown was induced at different levels, demonstrated excitotoxicity as well as dysregulated calcium signalling. The dysregulation of the analysed marker genes seen in both TSC1 models is consistent to that previously reported in both TSC and AD, strengthening the overlap between both disorders. Overall, both TSC1 cell models displayed significant dysregulation of markers throughout the development, demonstrating the importance of TSC1 expression for neuronal differentiation as acute loss caused significant dysregulation for weeks after the treatment. As these dysregulated markers align with observed AD pathology, this strengthens the hypothesis of that the TSC pathway will be an interesting target for future AD research and potential treatment.

This thesis advances deep learning models that are essential for collision avoidance of autonomous vehicles. The first contribution is an advanced multi-component loss function for 3D object detection algorithms where location of the object and dimensions of its bounding box are estimated simultaneously. The loss function penalises model's training process when the prediction does not match an expected ground truth. The proposed multi-component loss function enables to observe the progress of locating objects and place greater penalty on bounding box estimation when the object is well located and vice versa. This speeds up the training process as it helps the model to solve the easier task of locating the model first before solving the difficult problem of estimating its bounding box dimensions. Second, a novel sequential point cloud processing method for semantic segmentation is proposed. This uses a sequence of point clouds to generate a prediction. However, as point cloud processing is computationally expensive, processing sequences makes it even more computationally expensive. The proposed method alleviates this problem by fusing point cloud data in a latent feature space instead of processing all point clouds in the sequence each time a new prediction is made. As a result, the method takes advantage of sequential processing while keeping the computational overhead low. Finally, a practical unsupervised method to detect potential collisions in unlabelled point clouds is proposed. The method allows to test the performance and efficiency of different deep learning models on novel data without having to annotate the data first. It is based on the observation that most potential collision areas are defined by the closest object of interest (e.g., a car, a person). Also, the method provides a more realistic assessment of collision probability than widely used aggregate metrics.

Synchronization of autonomous oscillators is a phenomenon in natural and engineering systems with numerous applications. Biological synchronization is an area of particular interest owing to the added complexity of stochasticity inevitable in natural processes. This thesis focuses on physiological synchronization between cardiac rhythm and associated systems _ the respiratory system and locomotion. A novel experimental methodology has been developed with the specific aim of instigating cardiorespiratory synchronization at a range of breathing rates close to resting heart rate. All volunteers in this setting demonstrated at least one episode of 1:1 synchronization _ the strongest interaction for coupled systems. Synchronization for spontaneous breathing, as well as guided regimes, was also analysed using the same experimental protocol. Duration of synchronization was greater during guided breathing intervals when compared to spontaneous breathing, perhaps due to the reduction in variability inherent to breathing rhythm. Low breathing rates, known to maximise respiratory sinus arrhythmia (RSA), showed a lack of synchronization even when guided, supporting previous research suggesting that synchronization and RSA are opposing processes. Differences observed between athletes and non-athletes exhibiting 1:1 synchronization suggested a significant role of mechanical receptors within the cardiorespiratory interaction and invited the question of synchronization occurring during exercise. Running was considered as a third oscillatory process for an experimental protocol designed with breaths and steps controlled. Synchronization between the cardiac system and locomotion was observed, highlighting the interconnected nature of processes during rhythmic activity. Overall, the observation of 1:1 cardiorespiratory synchronization, in addition to synchronization occurring at spontaneous breathing rates - both guided and not - and during running, emphasises an additional component of connectivity within the cardiorespiratory system. Most likely, this component is linked to mechanical action in the heart. Furthermore, the consistency in results observed in the conducted experiments opens an avenue for further exploration of synchronization between physiological processes to understand the origin and implications of the synchronization phenomenon.

Scientific research has a profoundly important impact on our society and the environment. However, the multifaceted nature of this impact makes it particularly difficult to measure and, as shown in this thesis, it cannot be measured using traditional academic impact metrics that focus on counting citations and publications. Furthermore, existing societal and environmental impact metrics are only applicable to one scientific discipline or geography or are expensive processes run irregularly by government agencies. This thesis investigates natural language processing methods for identifying and measuring societal and environmental scientific impact and how such impact is reported in the news. A novel regression task and model are presented for identifying and quantifying this impact based on text extracted from scientific papers and news articles that discuss them. This is enabled by developing methods for linking and comparing news articles with academic papers that they discuss, whilst accounting for the structural and linguistic differences between the two types of document. Text encoding strategies for representation and comparison of long documents are also a focus of the thesis. A new cross-domain, co-reference resolution task between news articles and scientific papers is introduced so that co-referring entities may be used as anchors for aligning the two types of documents. Through comparisons of news article excerpts and sentences from corresponding scientific papers, it is shown that scientific discourse structure and argumentation in scientific papers is a likely predictor of which information will be presented prominently in news articles. This work introduces several novel natural language task settings for which no pre existing data sets exist. This has necessitated the production of new human-annotated datasets which were built using bespoke annotation tools that use semi-supervised learning to accelerate the labelling process and minimise the cognitive load of the task on the annotator. The thesis also makes use of low resource approaches including few-shot and multi-task learning to facilitate the development of accurate models with small data-sets. The resulting annotated data-sets, annotation tools and guidelines along with state-of-the art machine learning models are all made available as open assets. This thesis contributes new ways to measure societal and environmental impact of scientific work and help scientists and funding bodies understand how work is being used by others, justify the spending of public funding and inform better public engagement.

Learning audio representations is an important task with many potential applications. Whether it takes the shape of speech, music, or ambient sounds, audio is a common form of data that may communicate rich information. Audio representation learning is also a fundamental ingredient of deep learning. However, learning a good representation is a challenging task. Audio representation learning can also enable more accurate downstream tasks both in audio and video, such as emotion recognition. For audio representation learning, such a representation should contain the information needed to understand the input sound and make discriminative patterns. This necessitates a sizable volume of carefully annotated data, which requires a considerable amount of labour. In this thesis, we propose a set of models for audio representation learning. We address the discriminative patterns by proposing graph structure and graph neural network to further process it. Our work is the first to consider the graph structure for audio data. In contrast to existing methods that use approximation, our first model proposes a manual graph structure and uses a graph convolution layer with accurate graph convolution operation. In the second model, By integrating a graph inception network, we expand the manually created graph structure and simultaneously learn it with the primary objective in our model. In the third model, we addressed the dearth of annotated data by including a semi-supervised graph technique that represents audio corpora as nodes in a graph and connects them depending on label information in smaller subgraphs. We brought up the issue of leveraging multimodal data to improve audio representation learning in addition to earlier works. To accommodate multimodal input data, we included heterogeneous graph data to our fourth model. Additionally, we created a new graph architecture to handle multimodal data.

Treatment resistance is one of the key contributors to low overall survival rates in Colorectal cancer (CRC) patients. Although both resistance to chemotherapy and radiotherapy is observed in CRC patients, radioresistance of rectal cancer patients is a major issue in the clinic. 40% of patients with rectal cancer that receive neo-adjuvant radiotherapy will have no significant response to treatment, whereas only 10% of patients will have a complete pathological response (pathCR) to treatment. In order to understand pathCR radiosensitivity markers need to be defined. Studying radiotherapy with current models has been challenging due to the nature of the available models and their limitations to recapitulate the patients' tumour biology. The emerging new 3 dimensional (3D) organoid models show promise in regards to the modelling of cancer and patients' response with in vitro experiments that will be able to recapitulate tumour microenvironment. The aim of this study was to establish an organoid model for studying irradiation in rectal cancer patients and to identify radiosensitivity and radioresistance drivers in order to understand and tackle the problem of resistance to radiotherapy in clinical patients. Organoid lines were established and characterised for pathological and molecular features. The organoids were subjected to short-course radiotherapy (25Gy dose delivered in 5 fractions over the course of 5 days) and the response to the treatment was measured. Furthermore, the changes caused by the irradiation were investigated by performing whole genome sequencing, DNA methylation arrays, total RNA sequencing, and single-cell sequencing on irradiated and control organoids. Lastly, combination drug and radiotherapy assays were performed using organoids with mTOR and Akt inhibitors in order to sensitise cells to irradiation. The results revealed that irradiation causes changes on a genome-wide scale and disrupts the mTOR/PiK3CA signalling pathway. Combination therapy showed that Rapamycin is not effective in sensitising cells to irradiation, whereas AZD2014 was able to sensitise certain organoid lines to irradiation; the same was found to be the case for MK-2206. In conclusion, the results showed organoids pose as representative models for modelling radiotherapy response, and that blocking mTOR via dual inhibition of mTORC1 and mTORC2 as well as inhibition of Akt can sensitise cells to irradiation. Finally it was found that, in order to sensitise resistant lines, the dual inhibition of mTOR/Akt might be required.

"What I do not understand, I can still create."- H. Sayama. The following work follows closely the aforementioned bonmot. Guided by questions such as: "How can evolutionary processes exhibit learning behavior and consolidate knowledge?", "What are cognitive models of problem-solving?" and "How can we harness these altogether as computational techniques?", we clarify within this work essentials required to implement them for metaheuristic search and optimization. We therefore look into existing models of computational problem-solvers and compare these with existing methodology in literature. Particularly, we find that the meta-learning model, which frames problem-solving in terms of domain-specific inductive biases and the arbitration thereof through means of high-level abstractions resolves outstanding issues with methodology proposed within the literature. Noteworthy, it can be also related to ongoing research on algorithm selection and configuration frameworks. We therefore look in what it means to implement such a model by first identifying inductive biases in terms of algorithm components and modeling these with density estimation techniques. And secondly, propose methodology to process metadata generated by optimization algorithms in an automated manner through means of deep pattern recognition architectures for spatio-temporal feature extraction. At last we look into an exemplary shape optimization problem which allows us to gain insight into what it means to apply our methodology to application scenarios. We end our work with a discussion on future possible directions to explore and discuss the limitations of such frameworks for system deployment.

Female mosquitoes of the genus Anopheles transmit the parasites responsible for malaria, one of the most serious vector-borne diseases, which affects one-fifth of the world population. Current malaria prevention relies heavily on vector control interventions, which are most effective and sustainable when based on mosquito behaviour. Although much research has focussed on host-seeking behaviour, less is known about the close-range phase, including landing and post-landing behaviour. Thus, this research aimed to 1) quantify the relative effect that host-associated stimuli have on landing response, 2) quantify the effect that physical target characteristics have on flight behaviour and landing response, and 3) characterise pre- and post-landing behaviour and feeding success in relation to variations of target temperatures. Using behavioural assays, this project firstly quantified the relative role of three host-associated stimuli (thermal, visual, and olfactory) in driving landing and found that they act synergically to increase landing response, that host odour is an essential cue in this phase, and that the landing response is the result of a flexible yet accurate stimuli integration. Secondly, mosquitoes were offered targets with different physical characteristics, and it was found that more mosquitoes landed on the target when at least half of its surface was heated, which suggests a basis for more cost-effective trap designs. Furthermore, more mosquitoes landed on large targets, although small targets caught a greater density of mosquitoes per unit area; however, no effect on landing was observed when targets were oriented vertically or horizontally. Thirdly, it was demonstrated that the surface temperature of a target strongly influenced the feeding success, with mosquitoes being equally successful in feeding on blood at temperatures that ranged from normal human physiological to febrile conditions. Surface temperature also influenced post-landing behaviour, as mosquitoes displayed few foraging events and fed promptly on blood at 36 °C, whilst behaved radically different when presented with blood at 30 °C and 48 °C. Altogether, these results indicate that mosquitoes respond with distinct behaviours depending on the set of cues encountered during host-seeking. The results presented here could be used to improve vector surveillance and control tools by incorporating specific stimuli that are known to elicit specific behavioural responses. Thus, this research aids the cause of reducing the global burden of mosquito-transmitted diseases.

The Mountain Chicken Frog (Leptodactylus fallax, MCF) is a giant amphibian, endemic to the Lesser Antilles. Once an abundant wild food resource on its home islands of Dominica and Montserrat, the MCF became Critically Endangered and illegal to harvest in the 2000's following an outbreak of the infectious disease chytridiomycosis. Chytridiomycosis is caused by a fungal pathogen,Batrachochytrium dendrobatidis (Bd). Bd has driven population declines in over 500 amphibian species and is a major threat to amphibian biodiversity and the MCF's survival. Susceptibility to Bd infection can vary within amphibian host species, across habitats, and through time. Indeed, a small MCF population recovery was observed in 2014, with some MCFs even able to survive or clear Bd infection. The mechanisms enabling the survival of this population and its apparent tolerance to Bd infection are unclear, but also crucial to understand in order to conserve the MCF and aid wider amphibian conservation efforts. Changes in infectious disease dynamics occur with changes in one or more of: the host, the pathogen, and their environment. This thesis presents a series of investigations into potential host- and pathogen-related determinants of the MCF's susceptibility to or survival from the Bd epizootic, complementing existing studies which focus on environmental determinants. The first two host-related studies in this thesis involve a reconstruction of the demographic history of the MCF and a characterisation of key immune loci in the major histocompatibility complex (MHC). Historic MCF population reductions, identified by the demographic models, probably led to the depleted MHC diversity identified by the immunogenetic study. Low MHC diversity potentially explains the susceptibility of most MCFs to lethal Bd infection. Meanwhile, three MHC-I alleles were identified at an increased in frequency in Bd-tolerant MCFs, linking MHC-I diversity to Bd-tolerance for the first time. The increased frequency of these alleles in Dominica also explains why more MCFs survived there than in Montserrat. This thesis then carried out a novel characterisation of the MCF's skin microbiome, a key innate immune system component. This study revealed that the skin microbiome of surviving wild MCFs has potentially Bd-protective features, but these features were not explicitly tested against Bd infection. Replenishing microbiome diversity in Bd-naïve MCFs did not protect them from Bd infection, raising doubts around the utility of the skin microbiome in determining MCF survival from Bd infection. The last host-related study quantified remaining genetic diversity in captive MCFs and produced breeding recommendations which will influence future population recovery dynamics in the MCF. A final study into pathogen changes across the Bd epizootic revealed that the same, hypervirulent Bd lineage was implicated in the MCF's decline. As such, determinants of MCF survival or susceptibility to Bd infection are likely host- rather than pathogen-related. Following these studies, new avenues for research have been opened which will improve our understanding of amphibian chytridiomycosis. Conservation strategy is now more fully informed and better able to work towards mitigating Bd in the wild, a fundamental step in recovering the MCF.

Huntington's disease (HD) is a neurodegenerative disorder caused by a dominantly inherited CAG repeat expansion in the huntingtin gene (HTT). Multiple studies indicate an involvement of neuroinflammation in HD pathology, including several PET studies demonstrating that microglial activation correlates with disease severity in HD patients. Nonetheless, an open question is whether mutant HTT expression leads to cell-autonomous transcriptional and functional changes in microglia, which potentially contribute to the disease pathology. In this thesis, a patient-derived induced pluripotent stem cell (iPSC) model of HD with 109 CAG repeats (Q109), and isogenic controls with a corrected wild-type length of 22 repeats (Q22) were used. Q109 and Q22 iPSC were differentiated to microglia-like cells, as well as astrocytes and neurons. Cell types were characterised using immunocytochemistry as well as flow cytometry. Investigation of CAG repeat size was performed using PCR and capillary electrophoresis. Microglia function was investigated in monoculture using a wide range of assays as well as immunostaining of cellular compartments. Bulk RNA sequencing was performed on microglia-like cells under basal conditions and after pro-inflammatory stimulation. The secretome of microglia-like cell monocultures as well as microglia-astrocyte co-cultures was investigated via mass spectrometry. Both, Q109 and Q22 iPSC successfully differentiated to microglia-like cells, astrocytes and neurons. Q109 microglia-like cells, astrocytes and neurons demonstrated HTT CAG repeat expansion in culture. Q109 microglia-like cells showed impairment in key microglia functions, including phagocytosis and endocytosis and these functional impairments were also evident on a gene expression level. Furthermore, the secretion of pro-inflammatory cytokines was increased in Q109 microglia-like cells. RNA sequencing revealed an upregulation of genes involved in translation and mismatch repair in the comparison of Q109 versus Q22 microglia-like cells. In the investigation of microglia-astrocytes co-culture supernatants, astrocytes played a more prominent role for differentially secreted proteins. These observations demonstrate that iPSC-derived microglia-like cells can be utilized to better understand the role of microglia in HD pathology. The findings indicate a mutant HTT induced change in microglia transcription and cellular function which is independent of the detrimental HD brain environment.

Genetics plays a significant role in Alzheimer's disease (AD) risk, with multiple genetic risk variants uncovered by Genome Wide Association Studies (GWAS), many of which are enriched in microglia, the brains resident immune cells. Microglia are responsible for central nervous system (CNS) maintenance, injury response and fighting infection. There is clear evidence microglial dysfunction plays an critical role in AD pathogenesis, suggesting therapeutically targeting these cells may be of benefit. Our understanding of how these risk genes influence glial responses throughout age is limited, and how their activity in glia contributes to disease onset and development is largely unknown. This present study sets out to investigate potential glial roles of conserved AD risk genes, using Drosophila melanogaster as a model, with the aim to elucidate novel mechanisms contributing to disease risk. A reverse candidate knockdown screen of multiple conserved AD risk genes was conducted to assess AD risk genes roles in glial mediated neurological dysfunction and longevity. Results from this screen identified novel candidate genes, including orthologs of MEF2C, NME8 and ACE that expression in glia likely contributes to a healthy ageing nervous system and survival. PLCG2 is an important candidate gene to study AD relevant functions in flies, with a coding variant (P522R) linked to reduced late onset AD (LOAD) risk. This thesis aimed to characterise the glial role of the Drosophila PLCG2 ortholog, small wing and it substrates (PI(4,5)P2 and PI(3,4,5)P3) in ageing and Aβ42 related pathology. This study characterised the utility of genetically encoded, cell type specific fluorescent reporters for measuring the PI(4,5)P2 and PI(3,4,5)P3 dynamics in the fly brain, as well as a quantifiable model of glial driven Aβ42 toxicity. Finally, transgenic Drosophila expressing human PLCG2 wildtype (P522) and AD protective (R522) variants were created to explore functional changes associated with the hypermorphic R522 variant and its contribution to reduced LOAD risk. Results highlighted a conserved role for glial sl/PLCG2 in modifying Aβ42 toxicity and confirmed a protective role of the R522 reduced risk variant compared to the common P522 variant in Aβ42 associated pathology. Taken together these results indicate modulating PLCG2 activity may be a promising therapeutic target for treatment of AD.