42337 research outputs found
Sort by
Evaluating the Combined Edge Segment Texture (CEST) Method for Damage Detection in Al Fashir, Sudan
No full textThis study evaluates the applicability of the Combined Edge Segment Texture (CEST) method for detecting conflict-induced building damage in Al Fashir, Sudan, using moderate-resolution PlanetScope imagery. The research adapts a methodology originally designed for very high-resolution (VHR) data, implementing it through a hybrid workflow that combines Google Earth Engine and Python. The method integrates three analytical branches: edge detection, texture analysis, and segmentation which are then combined using a rule-based decision tree. The results reveal a fundamental incompatibility between the CEST method and the 3–5m resolution of the imagery, leading to a final F1-Score of 39/37%. The analysis found that the mixed-pixel effect and feature blending compromised each analytical branch, resulting in a low recall and a high rate of false positives. While the method lacked the precision for detailed, building-level assessments, it was sufficient at identifying major damage hotspots using visual approach. The findings suggest that the adapted CEST method serves as a tool for rapid, broad-scale damage mapping in a humanitarian context, offering a balance between the affordability of PlanetScope imagery and the need for timely information
Mapping the Maps: Implementing New Geographical Search Methods into the Royal Scottish Geographical Society’s Images for All Website
Get PDFThis research paper examine how access to digital map collections can be made more accurate and user-friendly. Using the Royal Scottish Geographical Society as a case study, it details how new geographical search methods for browsing the RSGS collection have been implemented into the Images for All website. It details the results of system and user requirements analyses carried out on the existing IFA website and MID Database, highlighting the importance of making access to the collections more user friendly. The existing text search has been improved with fuzzy search, Soundex and relevancy scoring and a user interface which allows users to toggle between ambiguous places has been added. In order to facilitate easier and geographically based searching, a clickable map that allows users to search the collection has been created. A ‘Featured Items’ page has been set up to promote the Society’s Collection
Molecular and temporal underpinnings of non-cell autonomous pathology in amyotrophic lateral sclerosis
No full textAmyotrophic Lateral Sclerosis (ALS) is an incurable and rapidly progressing neurodegenerative disease with limited treatment and diagnostic options. These constraints result in diagnostic delays, hindering early intervention. These theragnostic challenges are indicative of our limited understanding of the fundamental pathogenic mechanisms of ALS, especially during the initial stages of the disease. Here I propose to understand the temporal and non-cell autonomous molecular drivers of disease to address this gap in knowledge. ALS is a genetically, pathologically, and clinically heterogeneous disease, making molecular markers essential for clinical trial stratification and target engagement. Whilst most ALS cases are sporadic (sALS), approximately 10% are familial (fALS). Familial ALS cases implicate over 20 genes in the pathogenesis of the disease, with the most common being C9orf72. Although prior research has primarily focused on the impact of ALS on neuronal cells, emerging evidence indicates that non-cell-autonomous effects of non-neuronal cells play a significant role in ALS pathogenesis. Evidence suggests that neuroinflammatory astrogliosis is a common pathology in ALS, with studies showing that modulation of gliosis in mouse models can prolong lifespan. However, the contribution of glial pathology to ALS pathogenesis and the temporal nature of this contribution are poorly understood.
To address this knowledge gap, I have employed three interconnected approaches. First, I re-stratified existing datasets based on the extent of glial pathology. Specifically, I compared ALS cases (both fALS and sALS) categorized by the presence or absence of glial pathology. This approach aimed to generate data that highlight the impact of glial burden on pathology, potentially revealing theragnostic targets obscured by the inherent heterogeneity when comparing healthy controls and ALS cases. Secondly, I developed novel tools to accurately probe TDP-43 pathology in these cases. To do this, I used specific TDP-43 markers that are indicative of loss-of-function (cryptic exons) and gain-of-function (TDP-43 aptamers) to stratify cases. Finally , these tools were also used to link my findings to temporally stratify induced pluripotent stem cell (iPSC)-based models of fALS models. This approach aimed to enhance our comprehension of the temporal progression of pathogenic mechanisms in ALS. By leveraging these models, we can pinpoint and characterise the earliest stages of disease pathogenesis during critical developmental milestones, documenting initial disease-associated alterations and key molecular changes.
The findings provided compelling evidence that astrocytic TDP-43 pathology significantly associates with genetic and regional vulnerability, and molecular phenotype variability, suggesting that astrocytes actively contribute to disease progression through both gain- and loss-of-function mechanisms. Specifically, the inclusion of cryptic exons, indicative of TDP-43 nuclear loss of function, further confirmed this notion. Additionally, a systematic review of the literature revealed significant astrocyte involvement in ALS pathogenesis, demonstrating that dysfunctional astrocytes induce neurotoxicity and substantially affect motor neuron survival. Using patient-derived iPSC models, it was established that early pathogenic features, including altered TDP-43 and normal STMN2 expression, manifest early in the differentiation process, providing robust platforms for temporal mapping and therapeutic screening. Moreover, by employing NanoString sequencing in key developmental stages of motor neuron development, genes governing early neurodevelopment, including those involved in progenitor proliferation, axon elongation, and RNA metabolism, were downregulated in C9orf72 hiPSCs-derived cells relative to controls. Additionally, overlapping molecular signatures from post-mortem ALS tissues and hiPSC-derived models highlighted dysregulation in pathways related to myelination, metabolism, and neuroinflammation. Finally, novel single-cell transcriptomics methods developed during this research, including optimized hydrogel droplet sequencing (Hydrops), allowed for the capture of previously undetectable heterogeneity within ALS samples, thus greatly enhancing the resolution of pathological and therapeutic insights.
This thesis significantly advances our understanding of ALS pathogenesis by defining the temporal and cell-specific contributions of TDP-43 dysfunction and glial pathology, highlighting astrocyte involvement, and offering methodological innovations. These insights provide a foundation for developing targeted early-stage interventions and robust stratification strategies for future clinical trials, ultimately driving progress toward more personalised and effective ALS therapies
'Capitalising on the workforce': human capital valuation and the metrification of human resource management
Get PDFThe purpose of this thesis is to provide a critical analysis of the financialisation of human resources through an examination of Human Capital (HC). This thesis builds on the analysis of major institutional initiatives promoting HC reporting and emerging forms of HR management featuring intensifying metrification of HC and monitoring of employees. It links this managerial trend with the conception of value of the workforce derived from the changes in financialised capitalism. From a valuation perspective, this thesis situates the notion of value amongst the rhetorical apparatus that constitutes its intelligibility (theories that define what “true value” is) and the economic reality formed accordingly (operations that ascertain such value). The value of HC, in this context, is not intrinsic but produced by actors who promote it and results from practical valuation activities.
Human capital as a prevalent concept has been increasingly recognised in financial reporting and management, but its validity in financial valuation remains contested and its measurement is dominated by non-financial metrics. Based on a comparative analysis of organisational documents of several initiatives to promote HC across governments, transnational agencies and professional associations, as well as interviews with key actors who are directly involved in designing and measuring HC metrics, the research shows that the emphasis in the conceptions of HC differs across scholars and practitioners involved as key actors in shaping and promoting HC, depending on whether they are located in finance and accounting, management, or whether they are investors. Drawing on a growing body of literature in the economics and sociology of conventions (EC/SC) and valuation studies, I argue that the valuation of HC is situated in the coordination of three different conventions: the convention of financial accounting (shared by accounting professions), the convention of management accounting (shared by management professions) and the convention of financial value estimation (shared by investors). They have all come into play in the constitution of HC value and the way it is measured. The financialisation of HC is driven particularly by the convention of financial value estimation which conceptualises the value of HC from a financial market perspective through the “investor’s gaze”.
This study aims to contribute to existing scholarship in several aspects: 1) understanding the emerging trend of quantification and metrification in human resources management (HRM) from a valuation perspective, exploring the process of how HC as a concept derived from finance and economics is adopted in management and facilitates the financialisation of the valuation of the workforce, allowing the colonisation of financial valuation over non-financial metrics. 2) Providing a critical analysis of how the conception of HC is received and reacted to at organisational level, a level that has not been addressed much in existing scholarship where study in critical social sciences on HC often focus on the ideological and disciplinary effects of this idea over individuals. 3) Applying a pragmatist approach to develop a sociological critique of HC value in the research stream of economics and sociology of conventions (EC/SC) to reveal the configuration of HC and thus contribute to a better understanding of the financialisation of human resource
The use of metaphyseal sleeves to address bone loss in revision total knee arthroplasty
Get PDFThe rate of revision total knee arthroplasty is projected to increase by more than 300% between 2005 and 2030. This is due to a combination of ever-increasing number of primary procedures and increasing life expectancy of the population. Whilst survivorship of primary TKA has been well investigated, there has not yet been an ideal method for revision total knee arthroplasty with favourable long-term results. One of the most challenging aspects of revision total knee arthroplasty is managing bone loss, a number of techniques have been investigated with varying success, including augmentation with cement, bone graft and modular augmentation with blocks or wedges. This thesis aims to examine the use of metaphyseal sleeves through a series of retrospective, biomechanical and radiographical investigations.
A prospective database of patients undergoing metaphyseal sleeve revision total knee arthroplasty over a ten-year period was developed. This database was used to report on the outcomes of the surgery, the primary long-term outcomes were implant survival and infection rates. To further evaluate the performance of the metaphyseal sleeve, a radiographical analysis was carried out to compare stability of the implant. Osteolysis and Subsidence were examined on 100 sequential radiographs by 3 users. Inter and intra-observer variability was measured. A subgroup analysis was then performed to investigate if there was any difference in outcomes of patients who received a metaphyseal sleeve only versus a sleeve and intramedullary stem. Finally, a cadaveric biomechanical study was performed using nine cadaveric tibias with simulated increasing bone loss to assess this effect on stability.
For the retrospective study 120 metaphyseal sleeves were matched to 120 patients from the SAP database and 120 patients who underwent primary TKA in the same institution. Implant survivorship at 5 and 10 years was 98.3% for metaphyseal sleeves and 89.8% and 88.1% for the SAP control (p=0.0016). Metaphyseal sleeves had significantly higher overall implant survivorship across all time periods. There were 4 patients in the metaphyseal sleeve group and 5 in the SAP control group who had prosthetic joint infection (p=0.64). Oxford knee score was 33.4 for the metaphyseal sleeves and 35.7 for primary TKA. ROM was 98.1 degrees for the metaphyseal sleeves and 112.4 for the primary TKA. There was no significant difference in clinical outcomes, radiographic stability or complications when a metaphyseal sleeve is used with or without a stem for AORI type I and II bony defects.
Radiographical stability was investigated on 100 patients who received a metaphyseal sleeve revision TKA. The immediate post-operative radiograph was used as a baseline, and subsequent latest follow up radiograph used for analysis of osteolysis and implant subsidence. At final follow-up 3 patients (3.7%) had radiological subsidence of the prosthesis. 6 patients (7.5%) had radiological signs of lucency/osteolysis surrounding the implant. 1 patient required re-revision surgery (survivorship 98%). Metaphyseal sleeves provide excellent implant stability and survivorship up to 10 years. There is minimal subsidence, osteolysis and rates of revision.
In a cadaveric setting, increasing levels of bony defect were created in the tibia: small (15mm), medium (25mm) and large (35mm). Metaphyseal sleeves were inserted into the tibia and a standard axial pull-out (0 – 10mm) force and torque (0 - 30°) tests were carried out to determine the effect of increasing bone loss on implant primary stability. For the sawbone experiment the peak pull-out force with no defect was 160N, for small defect was 133N, medium defect 108N and large defect 56N. The peak torque with no defect was 17.8Nm, for small defect was 13.4Nm, medium defect 13.3Nm and large defect 11.1Nm. The only significant reduction of force was for the large defect. For the cadaveric experiment only the no defect, medium and large were tested. The mean peak pull-out force with no defect was 208N, medium defect 150N and large defect 50N. The peak torque with no defect was 13.6Nm, medium defect 9.9Nm and large defect 4.7Nm. The only significant reduction in force was observed in the large group. For the scenarios examined, this experimental study shows that when a metaphyseal sleeve is used for revision TKA, biomechanically up to 25mm radial defect may be acceptable, however between 26-35mm may be the “critical limit” for bone loss making the implant unstable. This is the first study in the literature to attempt to quantify this critical limit in any arthroplasty.
The findings of this thesis support the use of the metaphyseal sleeve for the management of bone loss in revision total knee arthroplasty. The failure rate, medium to long term patient reported outcome, radiographic evaluation and biomechanical properties are encouraging
Exploiting Klebsiella pneumoniae genomics for its significance in One Health
No full textThis thesis explores the genomics and epidemiological dynamics of Klebsiella variicola (K. var) and Klebsiella pneumoniae (K. pne), focusing on their adaptation to human hosts, resistance profiles, and virulence traits.
We begin by investigating the population structure and niche adaptation of K. var through a comprehensive genomic dataset. Our analysis reveals how this pathogen, traditionally associated with plants, has transitioned into the human niche through the acquisition of genes either due to antibiotic overuse or proximity to K. pne, highlighting its genetic variability and potential public health threat. To understand its behaviour in clinical infections, we conducted a case study in Scotland involving 119 clinical strains collected in 2023. This study identified multidrug-resistant and virulent strains carrying the yersiniabactin gene, underscoring the resistance and virulence of K. var in Scotland and emphasizing the need for ongoing surveillance.
We next shifted our focus to K. pne, a close relative of K. var, examining epidemic clones such as ST11, ST15, and ST258. We analysed their global dissemination and resistance/virulence profiles, with particular attention to ST11, which has caused significant outbreaks in China. Our genomic analysis revealed distinct evolutionary paths within the ST11 lineage, particularly in the Asian cluster, where resistance to extended-spectrum-β-lactams and virulence traits, such as aerobactin and yersiniabactin, were prevalent. Furthermore, we identified key resistance mechanisms linked to fluoroquinolone and colistin resistance, particularly in a hypervirulent clonal expansion cluster of ST11.
Overall, this thesis advances the understanding of how these Klebsiella species evolve, adapt, and contribute to significant public health challenges due to their resistance and virulence traits. These findings underscore the need for continued surveillance and control measures to mitigate the risks posed by these emerging pathogens under a One Health framework
Porting and Performance Tuning of SeisSol on Multiple HPC Architectures
Get PDFThe rising complexity of modern high-performance computing systems presents a significant challenge for scientific computing. Differences in hardware architectures and software environments require extensive, system-specific tuning to achieve high performance. As a result, these obstacles can make it
difficult for domain experts to run large-scale simulations efficiently, limiting the pace and scope of scientific research and discovery. In this work, we explore these challenges through a case study of SeisSol, an earthquake simulation code, by porting and tuning its performance across three different HPC systems. On the Bridges-2 system, as part of the Student Cluster Competition, we improved simulation performance by tuning the hybrid parallelism configuration and replacing suboptimal communication libraries. On ARCHER2, we further examined optimal hybrid parallelism strategies and identified a major bottleneck during the initialisation phase at scale. This was resolved by replacing the default communication backend, resulting in a substantial reduction in time-to-solution. Finally, on the GPU-based TeamEPCC cluster, we ported and addressed performance variability by updating a key communication library. These findings reinforce that performance is not automatically portable, and that architecture-aware tuning is essential for achieving scalable and efficient simulations across heterogeneous HPC environments
The use of smoking cessation applications in China: a realist evaluation
No full textBACKGROUND:
The epidemic of tobacco use has posed huge challenges to global public health. Consequently, tobacco control is crucial in achieving the United Nations Sustainable Development Goal 3 (good health and well-being). Smoking, as one of the most prevalent forms of tobacco use, is a leading cause of preventable morbidity and mortality worldwide. One measure of tobacco control is launching smoking cessation programmes, which have been recognised as an effective method to support individuals in quitting smoking. Smoking cessation programmes can be delivered through various approaches, including mobile applications (apps). According to the most updated data of the World Health Organization in 2024, China is the world's largest producer and consumer of tobacco, with over 30% of worldwide smokers coming from China. Given the high penetration rate of mobile phones in China, smoking cessation apps have great potential to assist Chinese smokers to stop smoking. Although the effectiveness of mobile health interventions in smoking cessation has been established by previous research, there is still a lack of research exploring how smoking cessation apps work.
AIM:
This study aims to examine what aspects of smoking cessation apps work for Chinese smokers, under what circumstances, and how.
METHODS:
The study employs the realist evaluation methodology and is composed of three stages. In stage one, two separate systematic reviews and six semi-structured interviews with Chinese health workers working in the field of smoking cessation were conducted to formulate initial programme theories (IPTs) in both the forms of ‘if...then...because...’ statements and Context-Mechanism-Outcome Configurations (CMOCs). In stage two, 24 semi-structured interviews were conducted with Chinese smokers who have experience using smoking cessation apps to test the IPTs. Stage three involved realist refinement of the IPTs leading to the development of refined programme theories, which proceeded by comparing the data with initial CMOCs deductively and identifying new constructs of programme theories inductively. The refined programme theories were presented in the form of CMOCs. Across this study, all semi-structured interview data were analysed using the thematic analysis approach. Ethics approval for the study was obtained from the University of Edinburgh.
FINDINGS:
In the context of being motivated to stop smoking and perceiving smoking cessation apps as a useful tool, smokers reported that engaging with the app features that tracked quitting achievements and health gains and showed the risks of smoking boosted their motivation and self-efficacy in smoking cessation, therefore supporting long-term abstinence.
Social features within apps provided crucial informational and emotional support, proving particularly beneficial for smokers seeking social support. Engaging with the social features enhanced their motivation, confidence, and self-esteem, although poor management and regulation of these social platforms within apps can lead to reduced engagement and affect the support necessary to quit smoking.
Although the withdrawal symptoms management features were intended to increase smokers’ skills to cope with withdrawal symptoms and clarify their craving patterns, the engagement level of these features was low. For those smokers who need the skills to prevent relapse and seek support within apps, the low intensity of intervention that lacks further guidance and personalisation may deter continuous use.
This evaluation also found that good usability and user experience could boost user engagement levels, therefore improving smoking cessation outcomes.
CONCLUSION:
This evaluation has generated explanations of the use of smoking cessation apps among Chinese smokers. The programme theories can be useful resources for app developers to design apps that better meet user needs to help them recover from addictive behaviours
Exploring the dark sector with higher-order statistics
Get PDFTo better understand the history of the dark matter distribution, we use
maps of the Universe to constrain cosmological parameters that describe the
characteristics of dark matter and dark energy. Various statistics calculated from
cosmological models have been used to measure and constrain parameters that
characterise different structures and effects in the Universe. Statistics measured
from gravitational lensing have been highly effective in probing the growth of
large-scale structures, as well as the accelerating expansion of the Universe.
Statistics measured from clustering and convergence maps can measure such
mass characteristics, which are then compared to theory predictions. Two-point
statistics are a powerful tool for constraining cosmology, with systematics and
measurement errors well understood. However, they can only measure signal from
Gaussian fields, where the collection of points have a Gaussian distribution. Even
though they do not capture non-Gaussian information on small-scale structures,
statistics involving two points have still been able to achieve precise constraints
when measured from weak lensing maps. To capture non-Gaussian information on
small-scales, I go beyond two-point and measure higher-order statistics in search
of stronger constraints on cosmological parameters.
In my first chapter, I investigate the inclusion of clustering maps in a weak
lensing Minkowski Functional (MF) analysis of Stage III and Stage IV survey
simulations to constrain cosmological parameters. The standard 3x2pt approach
to convergence and clustering data uses two-point correlations as its primary
statistic; MFs, morphological statistics describing the shape of matter fields,
provide additional information for non-Gaussian fields. Previous analyses have
studied MFs of lensing convergence maps; in this chapter I explore their
simultaneous application to clustering maps. I employ a simplified linear galaxy
bias model, and using a lognormal curved sky measurement and Monte Carlo
Markov Chain (MCMC) sampling process for parameter inference, I find that
MFs do not capture any additional signal compared to a 3x2pt analysis. However,
I argue that MFs should improve constraining power when nonlinear baryonic
and other small-scale effects are taken into account. As with a 3x2pt analysis,
I find a significant improvement to constraints when adding clustering data to
MF-only and MF+3x2pt weak lensing measurements, and strongly recommend
future higher order statistics be measured from both convergence and clustering
maps.
The process of converting cosmic shear maps to convergence maps involves tedious
calculations. Various mass mapping reconstruction methods have been developed
to simplify the process. In the second chapter, I evaluate the retention of signal
information of four reconstruction methods by measuring MFs from the output
maps. Kaiser-Squires (KS) takes the maximum likelihood estimate (MLE) of
the convergence map, applying Gaussian smoothing to reduce the noise. Wiener
Filter is a linear filter of signal and noise that applies a maximum a posteriori
(MAP) solution. Unlike KS, it can transform masked data, but neither are
able to capture non-Gaussian signal very well. DarkMappy approximates the
uncertainty that comes from reconstruction by reconstructing MAP convergence
maps. DeepMass is a deep learning model that specialises in denoising the signal.
It estimates the mean of the posterior distribution, where the iterative layered
network minimises the mean squared error (MSE) between the signal and the
estimate. In Chapter 3 I aim to determine which method best reconstructs the
original convergence signal with efficiency and accuracy.
LSST data will be even wider and deeper than any previous survey, providing
even more cosmological information with higher resolution weak lensing maps.
In Chapter 4, I explore the potential of higher-order statistics to extract
non-Gaussian information that is not accessible through traditional two-point
statistics. Specifically, I optimise the following higher-order statistics (HoS) for
an LSST framework: Minkowski functionals and peak counts, which decompose
the field at different scales. Measuring observables from noisy and masked LSSTlike
CosmoGridV1 simulations, I compare the information captured by the two-point
and higher-order statistics. Using a Fisher analysis to build contours, I assess the different statistics by forecasting their constraints on cosmological and
baryonic parameters, also studying the effects of survey geometry. I find both
HoS outperform two-point statistics, with MFs significantly improving precision.
I highly recommend the inclusion of Minkowski functionals in future higher-order
statistics analyses
Is IgA deficiency associated with canine atopic dermatitis?
Get PDFCanine atopic dermatitis (CAD) is a genetically predisposed skin disease characterised by inflammation and pruritus with an estimated prevalence of 3-15%. CAD is a complex syndrome, and despite extensive research, the aetiology is still not fully understood. Previous studies have shown that immune dysregulation, genetic predisposition, and skin barrier dysfunction all contribute to its pathogenesis. IgA deficiency (IgAD) has been associated with CAD in German Shepherd dogs though further investigation into whether this is the case in other breeds is required. The first objective of this study was to measure secretory IgA in faecal and saliva samples of atopic and healthy dogs from a range of breeds. Faecal sIgA concentrations were significantly lower in the atopic dogs compared to healthy controls (p value= 0.03). Samples in 22% of atopic dogs (n=7/32) and 34% of control dogs (n=12/35) fell below the suggested canine IgAD cut-off value (≤ 0.15 g/L). Human leukocyte antigen (HLA) genes have been associated with human atopic dermatitis, but less is known in canines. The second objective of this study was to examine dog leukocyte antigen (DLA) class II genes (DLA-DRB1, -DQA1 and -DQB1) in atopic dogs and breed-matched controls. DLA haplotypes and individual alleles were not associated with CAD in this analysis (p values >0.05), although the frequency of homozygosity in the case and control population was higher than anticipated (case=35% n=8/21, control=26% n=5/19). Skin barrier dysfunction can lead to secondary bacterial infections involving colonisation with Staphylococcus pseudintermedius. The third objective of this study was to examine Staphylococcal spp. mucosal carriage in dogs with CAD with and without dysbiosis. Pooled bacterial carriage samples (nares, axillary, inguinal and perianal) were taken from the atopic and healthy control dogs. If lesions were present on atopic dogs, samples of these were also taken. S. pseudintermedius isolates were identified in 30% (n=9/30) of the atopic population and 28% (n=15/54) of the control population. S. pseudintermedius isolates were identified in 59% of atopic lesion samples (n=10/17). Overall, the prevalence of S. pseudintermedius was lower than expected in case and control dogs
Access Repository Dashboard
Do you manage Edinburgh Research Archive? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!