ERA :: Home

ERA is a digital repository of original research produced at The University of Edinburgh. The archive contains documents written by, or affiliated with, academic authors, or units, based at Edinburgh that have sufficient quality to be collected and preserved by the Library, but which are not controlled by commercial publishers. Holdings include full-text digital doctoral theses, masters dissertations, project reports, briefing papers and out-of-print materials.

Information on current research activity including staff, projects and publications is available via the Edinburgh Research Explorer.

Now showing 1 - 40 of 41

Recent Submissions

Augmenting clinical risk prediction of cardiovascular disease through multi-omics
(2026-05-13) Chybowska, Aleksandra Daria; Marioni, Riccardo; Evans, Kathy; Price, Jackie; Medical Research Council (MRC); Medical Research Council Doctoral Training Program; University of Edinburgh: College of Medicine and Veterinary Medicine
BACKGROUND: The serum proteome can provide valuable insights into the development and progression of diseases. This is particularly important for cardiovascular disease (CVD), a leading cause of death worldwide. In this large-scale cohort study, we employ an untargeted massspectrometry- based approach to explore associations between highly expressed proteins, incident CVD (analysed as six individual outcomes and one composite outcome) and all-cause mortality. METHODS: The abundances of 439 proteins and protein groups quantified by mass spectrometry in serum were related to incident outcomes in 8,343 Generation Scotland participants (age 40–69 years), who were free of CVD at baseline (nall_cause_death=618, ncomposite_CVD=666, follow-up ≤17 years). Cox proportional hazards (PH) models were run before and after adjustment for preselected known CVD risk factors. Sex-specific effects were explored. A protein-based risk score for composite CVD outcome was developed using penalised regression. RESULTS: Forty-eight high abundance serum proteins and protein groups were significantly associated with incident CVD and death outcomes (PBonferroni<1.14x10-4), including 24 associations not reported in the Open Targets database. Proteins involved in immune and oxidative stress responses were associated with composite CVD (Immunoglobulin heavy variable 3/OR16-9, Hazard Ratio per SD (HR)=0.85 [95%CI 0.79,0.92]) and death (Alpha-1-antitrypsin, HR=1.27 [1.17, 1.38]), while heart failure was linked to proteins playing a role in lipid metabolism (Apolipoprotein A-II, HR=0.70 [0.59, 0.84]) and complement cascade (Complement C1q subcomponent subunit B, HR=1.40 [1.18, 1.66]). Applied to the test set, the proteomic risk score improved 17-year incident CVD prediction over models including age, sex, and nine lifestyle and clinical risk factors (ΔAUC = 0.010, ROC P = 0.013). CONCLUSION: The highly abundant serum proteome, readily assessed by mass spectrometry, reveals candidate biomarkers for incident CVD and provides predictive value for early risk stratification.
Desire, wealth and gender: anti-Buddhist prejudice in the Late Qing Dianshizhai Pictorial
(2026-05-13) Zhao, Yiming; Moore, Aaron; Huang, Xuelei
This research examines prejudice against Buddhists expressed in the late Qing publication Dianshizhai Pictorial and its underlying causes. As an influential and widely circulated pictorial of the era, the Pictorial exhibited a systematic bias in its portrayal of Buddhist monks and nuns. First, the pictorial depicted them as violators of monastic precepts, highlighting their inability to transcend worldly desires, with a particular focus on behaviors related to sexuality and wealth. Second, its criticism reveals a gender disparity: compared to monks who violated precepts of celibacy, nuns who committed similar transgressions received greater leniency. I argue that this stemmed from the nuns’ “liminal” status within the Confucian social order; their conduct both reflected the authors’ desires and anxieties, and reinforced beliefs in the necessity Confucian family ethics. Third, the substantial wealth of Buddhist clergy was subjected to widespread condemnation, closely linked to the frequent natural disasters of the late Qing and the quasi-public nature of temple assets. The Pictorial subjected Buddhist “unearned income” and use of funds to intense moral scrutiny. Finally, I suggest that public disillusionment with official military forces after the First Sino-Japanese War (1894-5) inspired an embrace of civilian martial traditions, symbolized by the Shaolin Temple. This shift represented a turning point that enabled Buddhists to begin reforming their public image during the late Qing and early Republican period.
Use of thermoplastic resins in composite tidal turbine blades: manufacturing feasibility and the effects of seawater immersion
(2026-05-12) Smith, Ione L. M.; Obande, Ogwa; Oterkus, Selda; Thies, Philipp; Alam, Parvez; Industrial CDT in Offshore Renewable Energy; FastBlade
Tidal energy is a predictable and reliable renewable energy resource with significant potential to contribute to a diverse and sustainable electricity mix. As integral components of tidal energy converters, tidal turbine blades operate in challenging conditions, where constant submersion in seawater and high, cyclic loads necessitate thick wall sections (10–100 mm), particularly at the root of the blade. Tidal turbine blades are typically manufactured from fibre-reinforced thermoset-matrix composites using vacuum infusion processing. Although thermoset matrices have traditionally been preferred for composite applications, the cross-links that provide their desirable mechanical properties also prevent melting and reshaping, limiting recyclability compared to thermoplastics. Increasing interest in sustainable alternatives has driven the development of recyclable, room-temperature processable liquid thermoplastic resins (LTPRs) to replace thermosets in vacuum infusion, offering the potential for more sustainable material solutions. To validate their suitability for tidal energy applications, it is essential to understand how these LTPRs perform during the manufacture of thick-section composites. Challenges typically associated with composite laminate manufacture are exacerbated in the production of thick-section laminates, including managing the thermal effects of polymerisation to avoid boiling and defects. It is therefore critical to understand the thermal profiles experienced during processing and the resulting laminate quality. Additionally, given the harsh marine operating environment, the gravimetric and microstructure effects of long-term submersion must be evaluated. This thesis investigates the suitability of LTPRs for the production of thick-section composites used in tidal turbine blades. Current blade manufacturing practices are reviewed, and alternative techniques explored to advance production methods. Two low-exotherm, acrylic-based LTPRs, Elium® 188 XO (E188 XO) and Elium® 191 XO/SA (E191 XO/SA), designed specifically for vacuum infusion, are considered. Thermal analysis studies using embedded thermocouples demonstrated that both LTPRs remained below their boiling points during the manufacture of 10-ply glass fibre-reinforced laminates. E188 XO exhibited a higher peak temperature and interlaminar thermal gradient but reached this peak more quickly, implying a higher manufacturing rate. Further studies with 20- and 30-ply laminates, produced using E188 XO, indicated that interlaminar temperature increases with laminate thickness. Nonetheless, peak temperatures did not exceed the resin’s boiling point, suggesting a minimised risk of exotherm-induced defects. These findings provide critical insights into polymerisation behaviour and support process optimisation. Variations of vacuum infusion processing were used to fabricate four E188 XO 52-ply laminates, with the baseline setup producing a laminate with a polymerised thickness of 42.5 mm. Surface temperatures varied between configurations, but remained below the boiling point of the resin. Notwithstanding, higher internal temperatures may have been reached, but external instrumentation enabled enhanced vacuum integrity during processing. Thickness variation was observed in all polymerised laminates. Fibre volume fraction and void content varied along the length and through the thickness of each laminate but were found to be relatively uniform across the width in most laminates. Of the four configurations assessed, notably lower void contents were observed in laminates produced with a semi-permeable membrane and the baseline setup. Quantitative results suggest a variation in laminate quality along the length as well as through the thickness of thick-section laminates. The long-term water uptake behaviour of the two LTPRs was assessed through an experimentally simulated 20-year service period. All 10-ply glass fibre-reinforced laminate samples reached saturation within this time. E191 XO/SA exhibited substantially faster water uptake, with considerable deviation in behaviour between samples. Samples extracted closer to the resin inlet end of the laminate exhibited rapid increases in mass that continued over time, while those from nearer the vacuum outlet showed a more gradual mass increase, followed by a steady decrease. Sample position from the laminate appeared to have a stronger influence on water uptake than delamination caused by cutting of the samples from the laminate. The variation in behaviour between samples from the same laminate is attributed to excessive voids towards the inlet end. Samples from E188 XO 20- and 30-ply laminates did not reach saturation, though inlet-end samples again showed higher water uptake and void content in the 30-ply laminate. These results highlight the challenges in thick-section manufacture and the complex relationships between materials, manufacturing, microstructure, and long-term behaviour in a marine environment. While variations in quality through the thickness are apparent, this work highlights that variations across laminates, including both length and thickness, should be carefully considered and managed during production. Overall, this thesis provides novel insights into the polymerisation behaviour, laminate quality, and water uptake performance of LTPRs in thick-section composites. These findings offer value to academics and industrial practitioners to support informed and strategic choices to improve and advance tidal turbine blade production.
Development of biosensor technology for the measurement and control of off-note flavours in the Scotch Whisky industry
(2026-05-12) Sabulyte , Julija; French, Chris; Laohakunakorn, Nadanai; Industrial Biotechnology Innovation Centre (IBioIC)
Flavour monitoring is routinely carried out during Scotch whisky production to track the development of the desirable flavour profile and for early detection of undesirable, off-note flavours. Standard analytical methods often fail to account for the effects of mixtures of compounds and may lack the required sensitivity, whereas human assessment has an unavoidable level of variability, thus, alternative approaches are needed. Recently, cell-based biosensors have emerged as cheap and portable sensory devices which repurpose the natural ability of living systems to sense and respond to changing conditions. Sensors based on live cells cannot be used outside a laboratory setting, but this can be overcome by the use of cell-free transcription-translation (TX/TL) systems, or even simpler transcription-only (TXO) systems, which are cheap and can give responses visible by eye within minutes. This project aimed to apply biosensor technology for the detection of whisky off-notes associated with 'musty', 'sour', 'solvent', and 'phenolic' flavours. First, the cell-free system was optimised for high transcriptional output and signal-to-noise ratio, low production cost and ease of use. Second, inducible promoters and transcription factors (TFs) responsive to the known off-note aroma compounds were identified through reports in the literature and transcriptomic analyses. And finally, alternative detection methods, such as through commercially available assays and off-note inhibited enzymic assays, were explored. Comparisons with sensory panel results suggested that while biosensor systems show promise, particularly in detecting certain sour and phenolic off-notes, the sensitivity of trained human assessors remains difficult to surpass. Nonetheless, continued development of biosensor technologies holds exciting potential to complement traditional flavour assessment, paving the way toward faster and more consistent flavour monitoring in the whisky industry.
Computational study of Beta-lactamases: from molecular dynamics to alchemistry
(2026-05-12) Güven, J. Jasmin; Mey, Antonia; Horrocks, Mathew
Quantifying the binding affinity of small molecules binding to protein targets is an important aspect of early-stage drug discovery pipelines. Computational methods are often used at different stages of these pipelines for increased efficiency and reduced costs. For example, alchemical relative binding free energy (RBFE) calculations can be used to achieve accurate predictions of binding free energies and can be used in the lead optimisation stage of early-stage drug discovery projects. However, this is only true for systems whose inter- and intramolecular interactions can be modelled accurately. Widely used in academia and industry, RBFE calculations can be easily applied to some protein targets, but there are many examples where the setup of these calculations is more involved. For example, open-source tools do not support the easy setup of RBFE calculations with metalloproteins, especially if the small molecules directly bind to the active-site metal. Given that nearly a third of proteins contain transition metals, developing accurate, open-source methods for parameterising metals to be used with alchemical free energy calculations is crucial. Thus, the aim of this thesis is to explore how current metal-modelling options can be used for molecular dynamics (MD) studies of systems where small molecules bind to metalloproteins. This includes standard MD protocols, as well as, the development of RBFE calculations for these systems, and their accuracy is evaluated against experimental results. While there are many pharmaceutically relevant metalloproteins, in this thesis we will focus on computational studies of β-lactamases, a class of enzymes responsible for the antibiotic resistance in most Gram-negative bacteria. Specifically, we will be focussing on the modelling of zinc coordination in a metallo-β-lactamase (MBL), a specific class of β-lactamases containing either one or two zinc ions in their active sites. Additionally, we explore how RBFE calculations between a reactant and a product may be used to guide future inhibitor design of a glycyl radical enzyme (GRE) – a strategy also applicable for future β-lactamase inhibitor design. Chapter 1 will introduce the role of β-lactamases in antibiotic resistance and how computational methods may be used to aid in drug discovery. In Chapter 2, we will briefly explain the theoretical foundations underlying MD simulations and alchemical RBFE calculations. In Chapter 3, we describe the structure of our python-based code workflow, meze, for parameterising metalloproteins for MD simulations and RBFE calculations. We will show minimal working examples of meze and show results from simulations prepared with this code in subsequent chapters. In Chapter 4, we use classical MD simulations of 16 experimentally found inhibitors bound to a serine-β-lactamase (SBL). This set of inhibitors, taken from literature, were found to inhibit this SBL and two MBLs. Exhibiting cross-class affinity is an important property of successful inhibition of β-lactamases, as many bacteria often produce multiple classes of these enzymes. Our MD simulations complement the experimental results by allowing us to analyse the interactions between the individual inhibitors and the SBL active site. These simulations highlight that the chosen simulation setup and force field choices model key non-covalent interactions adequately and lay the foundation for modelling MBLs with the same inhibitors, presented in Chapter 5. We use the meze tool to parameterise an MBL with two active-site zinc ions, using three different metal-modelling options for MD simulations. The first two models are so-called non-bonded models, where the metal coordination bond is not modelled directly: a restraint-based approach and an upgraded Amber force field (UAFF) for zinc-metalloproteins. The third metal model is a hybrid bonded model, where coordination bonds between zinc and coordinating residues, including water, are modelled with covalent bonds, while the bond between the inhibitor and zinc is removed. Our MD simulations with the two non-bonded models reveal that the experimentally observed tetrahedral coordination of both zincs is not conserved. The hybrid bonded model conserves the zinc coordination better, due to the explicitly defined coordination bonds, and more accurate partial charges, with the caveat of having to run expensive quantum mechanical calculations to refit the partial charges. In Chapter 6, we introduce how RBFE calculations can be used to study the relative binding affinity of the reactant and product of a GRE, responsible for converting choline into trimethylamine (TMA), which is further converted into trimethylamine-N-oxide (TMAO) in the liver. Both TMA and TMAO have been associated in multiple diseases, such as non-alcoholic fatty liver and cardiovascular diseases. The RBFE between these two molecules, and the differences in their interactions with the active site during these simulations may be exploited in future work for designing enzyme inhibitors. Finally, in Chapter 7, we show how well existing RBFE methods are able to reproduce experimental trends for β-lactamases. Conventional RBFE methods can be directly applied to SBLs, giving rise to expected literature accuracies for the ligand series. For the MBL, we use the two non-bonded models to carry out RBFE calculations and find that the UAFF approach improves correlation compared to the restraint approach. These results show that the same metal coordination issues present in the MD simulations also affect the accuracy of binding affinity predictions for the MBL. Furthermore, our results highlight that some key interactions, such as halogen bonding are not accurately modelled with the small molecule force field and may be the cause of outliers in RBFE calculations for both β-lactamases. In this thesis, we show that simple, non-bonded metal parameter models can provide predictive free energy estimates for a representative MBL. We outline a proof-of-concept metal-parameterisation tool, which aims to improve the ease of use of preparing metalloproteins for both MD simulations and RBFE calculations. We identify issues with the small molecule force field, and the metal models' ability to describe zinc-coordination sites, which may be reducing the accuracy of these methods. In Chapter 8, we outline these conclusions and include future directions into how metal modelling options could be improved for MD simulations and RBFE calculations.