22962 research outputs found
Sort by
Temporal effects of ovarian failure on skeletal muscle form and function: Insights from single fibre mechanics and histological analyses in a mouse model of gradual estrogen deficiency
No full textMenopause is associated with impairments and declines in skeletal muscle contractile performance and shifts in muscle composition, but the mechanisms and time course of these alterations are not well understood. Importantly, skeletal muscle adaptations to estrogen deficiency are modulated by metabolic status and mechanical loading, with diet and physical activity influencing the extent and nature of muscle remodeling. Most preclinical studies studying menopause use ovariectomy, which induces abrupt estrogen deficiency and fails to replicate the gradual menopausal transition in women. To overcome this limitation, we employed the 4-vinylcyclohexene diepoxide (VCD) model of ovarian failure in female CD-1 mice to investigate how progressive estrogen loss impacts skeletal muscle contractility and morphology across peri-, early-, and late-menopausal stages. (Study 1) Permeabilized single muscle fibres from the soleus (SOL) and extensor digitorum longus (EDL) were assessed for absolute and specific force, cross-sectional area, rate of force redevelopment (ktr), instantaneous stiffness, and calcium sensitivity. In the SOL, absolute force and fibre size were elevated following ovarian failure, accompanied by faster ktr and enhanced calcium sensitivity in type I fibres. However, specific force showed a biphasic pattern, with a transient decline during early menopause followed by recovery at later stages. In contrast, the EDL exhibited minimal changes, highlighting muscle-specific responsiveness to estrogen deficiency (Study 2). Histological analyses revealed that ovarian failure alone did not induce pathological remodeling of skeletal muscle (Study 3). Instead, modest, muscle-specific changes in intramuscular lipid content and fibre-type distribution emerged under defined metabolic or mechanical contexts, including high-fat diet or exercise exposure, while collagen content remained unchanged. These findings indicate that structural outcomes were shaped more strongly by metabolic context than by ovarian hormone loss per se. Thus, these findings demonstrate that gradual ovarian failure produces muscle-specific and time-dependent adaptations that differ fundamentally from the uniform deficits reported following ovariectomy. Intrinsic contractile disturbances can precede detectable morphological changes, and structural deterioration is not an inevitable consequence of estrogen loss in the absence of metabolic challenge. This work establishes the VCD model as a translationally relevant model for studying menopausal muscle biology and highlights the importance of metabolic context, mechanical loading and time course in shaping muscle quality during ovarian hormone decline.2026-12-1
DiffusionRank: A Denoising Diffusion Framework for Learning to Rank
Full text linkLearning-to-rank (LTR) is a cornerstone of modern information retrieval systems, yet most existing methods rely on deterministic scoring functions that struggle to capture the complex, high-order interactions among document features. In this thesis, we cast LTR as generative imputation and introduce DiffusionRank, the first diffusion-based framework for ranking. Specifically, we begin with perturbed input features and mask the relevance label, and through iterative denoising steps, the diffusion model learns to reconstruct both the input and the masked label. This formulation allows the model to capture the joint distribution over features and relevance labels, thereby learning complex feature-label interactions that are often overlooked in traditional discriminative models. Extensive experiments on three widely adopted benchmark collections, namely MQ2007, MQ2008, and MSLR-WEB10K, demonstrate the effectiveness of our approach. Without any external signals or architectural tricks, DiffusionRank achieves a 3% improvement in NDCG@10 over the discriminative point-wise baseline. In addition to the supervised setting, we address the practical limitation of data scarcity in real-world ranking applications, where annotating relevance labels is both time-consuming and labor-intensive. To this end, we propose a pre-train-then-fine-tune approach that exploits abundant unlabeled data. We pre-train the diffusion model on the full training set with the label reconstruction loss disabled, forcing the network to internalize the feature manifold. Subsequent fine-tuning with only a fraction of labeled data consistently outperforms fully-supervised baselines trained on the same label budget, underscoring the value of generative pre-training for data-efficient ranking. Overall, this thesis establishes diffusion modeling as a compelling, flexible, and label-efficient alternative for learning to rank, opening new avenues for generative relevance estimation in information retrieval
Development and validation of a molecular assay for the identification and enumeration of Eimeria species (Apicomplexa: Coccidia: Eimeriidae) infecting commercial poultry to supplement Oocyst-per-Gram counts
No full textCoccidiosis caused by Eimeria species commonly affects poultry production, impacting bird health and costing billions of dollars annually. Traditional oocyst per gram counts (OPG’s) provide reasonably accurate enumeration of oocysts and allow clinicians to estimate, indirectly, the severity of an infection and assign appropriate treatment but provide no information regarding relative numbers of constitutive Eimeria species. Morphological methods are less reliable for species identifications; flow cytometry and image analysis algorithms have been attempted but differentiating among numerous Eimeria species that infect domestic fowl becomes challenging. Molecular methods (e.g., ddPCR or deep sequencing (NGS) of amplicons from diagnostic loci) can provide accurate species identifications and quantifications but current approaches have limitations. This thesis outlines the development and validation of a probe-based qPCR diagnostic assay targeting a mitochondrial locus that adds parasite species differentiation and relative abundance to existing OPG counts providing accurate, economical and timely diagnostic information to the poultry industry.Natural Sciences and Engineering Research Council of CanadaOntario Ministry of Farming, Agriculture and Agribusiness and Ministry of Rural Affairs2026-12-0
Development of Compatibilization Approaches for Biodegradable Polymeric Blends and Biocomposites through Reactive Extrusion for Packaging Applications
No full textThis dissertation investigates compatibilization strategies to enhance the performance of biodegradable polymer blends and biocomposites, with a focus on reactive extrusion. The primary objective was to improve interfacial adhesion between immiscible biopolymers or to develop novel compatibilizers for biodegradable systems. The resulting materials were processed through injection molding, cast extrusion, and film blowing, and thoroughly characterized to assess performance improvements. The initial study examined in situ transesterification between two biodegradable polymers during melt processing. Formation of copolymers through this reaction acted as reactive compatibilizers, leading to a balanced combination of stiffness and ductility in PBSA/PBAT blends. A subsequent study involved PBSA-based binary and ternary blends to produce mono- and multilayer films. The use of organic peroxide during reactive extrusion enhanced melt strength and processing stability, yielding films with improved gas barriers and mechanical properties. Further research explored dual compatibilization using both an organic peroxide and a crosslinking agent. This dual approach significantly improved impact strength and elongation at break in PBSA-based blends, attributed to enhanced interfacial bonding. Another study assessed nanoclays as compatibilizers. Their dispersion and interfacial localization were highly dependent on processing conditions. A two-step extrusion process promoted nanoclay exfoliation and improved compatibility by localizing the nanoclays at the polymer interface. The final study developed a novel compatibilizer by grafting maleic anhydride onto cellulose acetate via reactive extrusion in the presence of an organic peroxide. The resulting maleated cellulose acetate exhibited strong potential as a compatibilizer in cellulose acetate-based blends. Overall, this research highlights reactive extrusion as a powerful tool for enhancing compatibility and performance in biodegradable polymer systems. The developed materials show strong potential for use in sustainable packaging applications, offering a viable alternative to conventional petroleum-based plastics.2026-11-2
“Let us be not only hearers but doers of the Word”: Affective Plasticity, Emotionology, and Embodied Responses of Scottish Covenanting Women*, 1638–1688.
Full text linkThis thesis argues that the Scottish Covenanting movement must be reinterpreted as an emotionally saturated regime in which women* negotiated, subverted, and reconstituted gendered norms through affective, emotional, and embodied practices of religion and resistance. Focusing on the period surrounding the National Covenant of 1638, the thesis examines how ministers and authoritative authors scripted emotional expression and confined women to passive, domestic piety. However, after 1638, new expectations of defiance and steadfastness enabled women to move from “hearers" to “doers," appropriating emotional styles coded as masculine and engaging in practices of self-degendering that unsettled patriarchal hierarchies of authority. Bridging affect theory, the history of emotions, and queer critique into conversation with early modern Scottish sources, this thesis challenges static narratives and is better able to foreground the lived experiences and emotional labour through which marginalized agents reconfigured their identities.Social Sciences and Humanities Research Council of CanadaUniversity of Guelp
Evaluating Source Code Embeddings from LLMs for Educational Downstream Tasks
Full text linkRecent advances in Deep Learning (DL) have transformed natural language processing by enabling models to capture complex linguistic patterns. These techniques extend naturally to source code, which shares similarities with natural language but also exhibits strict syntactic and semantic constraints. This thesis investigates the effectiveness of different source code embedding strategies for three educational tasks: code correctness, code summarization, and code similarity. We introduce a set of systematic evaluation frameworks and propose a fine-tuning approach for a pre-trained, code-centric large language model to generate context-aware C++ embeddings. Experimental results demonstrate that while modern embeddings capture meaningful semantic and functional information and perform well on summarization and similarity tasks, they struggle to encode fine-grained structural details required for reliable correctness assessment. These findings highlight key limitations of current embedding approaches and identify important challenges for automated grading and other educational applications that depend on precise structural understanding of source code
Fertilizer Nitrogen Dynamics in Inbred Seed Corn (Zea mays L.): Yield Response, Utilization and Residual Effects in Southwestern Ontario
Full text linkInbred seed corn (Zea mays L.) production in Southwestern Ontario benefits from favorable agro-climatic conditions but faces critical challenges in fertilizer nitrogen (FN) management. The objective of this dissertation was to establish a regional FN recommendation for inbred seed corn through accessing yield response, uptake, distribution and fate of FN in the soil-plant system using a three-study approach. The first study evaluated 46 inbreds from 2015 to 2018 across 13 site-years. Significant variability was observed, with 27 inbreds demonstrating quadratic marketable yield responses to FN, with an agronomic optimum N rate (AONR) ranging from 115 to 210 kg N ha⁻¹. When AONR data pooled by sites, regardless of responsiveness, an overall AONR was 162 kg N ha⁻¹, providing a regionally relevant benchmark recommendation for seed corn production in Southwestern Ontario (i.e. the Southwestern Ontario FN recommendation). However, applications of 180 kg N ha⁻¹ significantly increased residual soil inorganic N (70 kg N ha⁻¹), identifying a threshold for environmental risk. The second study used 15N-labeled FN in microplots on loamy sand and loam soils to track the fate of FN. Findings showed that FN uptake was more heavily influenced by soil type than inbreds, with grain N content being 22% higher in loamy sand than the loam soil. Notably, 25% of applied FN was immobilized in soil organic pool, while 12 to 30% was unaccounted for by harvest depending on soil type. The third study used residue exchange approach to assess the fate of residual FN. Results confirmed that residual FN is an insignificant N source, contributing only 1 to 3 kg N ha⁻¹ to spring wheat (Triticum aestivum L.) uptake, while up to 50 kg N ha⁻¹ was unaccounted for and presumed lost during the non-growing season. Together, these findings demonstrate that a single fixed FN rate cannot adequately account for the high variability in inbred response. To enhance both sustainability and productivity, future management should adopt the 4R Nutrient Stewardship framework to refine nitrogen applications
Proteomics of Regeneration in Scleractinian and Corallimorpharian Corals and During a Corallimorpharian Mortality Event
No full textCoral regeneration is a critical biological function for natural survival and supports asexual propagation in coral reef restoration efforts and in the saltwater aquarium industry. This study investigates the potential mechanisms underlying regeneration and survival of corallimorpharian corals (Ricordea florida, Discosoma sp.) using proteomics and light microscopy. It also includes a comparative proteomic analysis of regeneration in three reef-building, scleractinian corals (Montipora capricornis, Seriatopora hystrix, Pocillopora damicornis). An overarching goal of the thesis was to identify protein biomarkers for regeneration present in multiple coral species. The corallimorpharians R. florida and Discosoma sp. were bilaterally fragmented and regeneration rates, survival and differentially abundant (DA) proteins were analyzed. Polyp colour significantly (p = 0.001) affected the rate of regeneration, while tentacle size and polyp grouping (isolated polyp vs. close contact) did not. Bilaterally fragmented R. florida completed regeneration within 3-24 days post-fragmentation with over 83% survival. A comparison of fast vs slowly regenerating polyps identified 18 DA proteins linked to energy production and immune homeostasis in the faster-regenerating polyps. All scleractinian corals completed regeneration within 28 days with 100 % survival with smaller fragments regenerating faster, while larger fragments showed significant weight gain (p < 0.05). Thirteen DA proteins linked to stress response, apoptosis and reproduction were shared between M. capricornis and S. hystrix. By comparing DA proteins found in this thesis and three previous research efforts, five proteins, ATP synthase subunit beta mitochondrial, failed axon connections homolog and soma ferritin-like, serine-pyruvate aminotransferase-like and glyoxalase 3-like, were identified in multiple species and experiments, making them strong biomarker candidates for regeneration. During maintenance of corallimorphs used in this project, two unexpected adverse health events occurred, “Ricordea yuma wasting syndrome” and “stasis” of R. florida that provided insights into molecular mechanisms involved in these pathologies. Proteomics revealed DA proteins linked to apoptosis in R. yuma polyps with wasting syndrome, supported by light microscopy evidence of amoebocyte migration, although experimental replication of the syndrome was unsuccessful. These findings contribute to our knowledge of coral resilience and support improved propagation techniques in conservation and aquarium industry practices.Natural Sciences and Engineering Research Council of Canad
Optimizing Best Management Practices for In-season Nitrogen Applications in Corn (Zea mays L.) to Minimize Gaseous Nitrogen Losses and Enhance Yield
Full text linkMaintaining corn productivity while safeguarding environmental quality and human well-being remains a major challenge. Corn production relies on large nitrogen (N) inputs that frequently result in ammonia (NH₃) and nitrous oxide (N₂O) losses. Identifying N management strategies that reduce these losses without compromising yield is therefore critical. The thesis is comprised of four studies. First, a modelling study using the calibrated Denitrification and Decomposition (DNDC) model evaluated the long-term (39-year) effects of N application timing (at-planting vs. in-season at V6) across multiple planting dates (PDs). Large single in-season N applications resulted in higher N₂O emissions than at-planting applications. PD moderated this effect, with larger differences observed at earlier PDs. Reduced emissions from at-planting N at early PDs were primarily driven by lower soil temperatures that suppressed microbial denitrification. Second, the Dosi-tube method originally developed to measure NH₃ volatilization from manure was modified to quantify volatilization from inorganic N fertilizers under a corn canopy. The revised method improved accuracy relative to the original algorithm, providing a simple and cost-effective approach for on-farm volatilization measurements. Third, a multi-environment field study assessed the effects of N rate, timing, source (urea vs. urea–ammonium–nitrate [UAN], with or without urease inhibitors [UI]), and placement on NH₃ volatilization and grain yield. Injection or UI use substantially reduced NH₃ losses compared with unprotected surface-applied urea, particularly at V13. However, yield gains from reduced volatilization occurred only in environments where crop N demand exceeded total N supply. Soil cation exchange capacity (CEC) and rainfall dynamics were identified as key environmental drivers of volatilization risk at V13. Finally, the agronomic consequences of late-season N applications (>V10) were evaluated. Holding the total N rate constant (179 kg N ha-1), different proportions of N were applied late in-season at V13 (80%, 50% and 30% applied in-season) with the remainder applied around planting. Increasing the proportion of N applied at V13 generally reduced yield due to diminished vegetative N status and greater reliance on post-R1 N uptake. This penalty was mitigated under conditions favoring post-R1 N acquisition, such as coarse-textured soils.Agriculture and Agri-Food CanadaOntario Agri-Food Innovation AllianceGrain Farmers of OntarioNatural Sciences and Engineering Research Council of Canad
Multi-Sensor Earth Observations for Monitoring the Dynamics and Water Surface Elevation of Arctic-Boreal Lakes
No full textCharacterizing and quantifying Arctic lake dynamics is essential, as they are a critical component of the global hydrological cycle, and play a crucial role in regulating greenhouse gases, supporting aquatic organisms, and influencing other ecological functions within the Arctic-Boreal region. However, challenges such as the region's remoteness, the high cost of maintaining in-situ instruments, and seasonal freezing limit their effective monitoring. To bridge these gaps, this PhD thesis enhances the tracking of two key processes that underpin lake dynamics using multi-sensor earth observation instruments data. The first research chapter examined four decades (1984-2024) of spatial distribution of North America Arctic-Boreal lake drainage and formation using Landsat time series, while also identifying the hydro-climate processes driving these changes. Out of the total 1.5 million detected lakes, approximately 874 drained lakes (207.121 km²) are associated with a rise in air temperature, along with changes in snow water equivalent and surface runoff. About 596 formed lakes (175.6 km²) were also influenced by warming air temperatures. The second research chapter leveraged the complementary strengths of multiple altimetry satellites to retrieve water surface elevation (WSE) of thermokarst lakes. Specifically, the temporal advantage of Sentinel-3 and the spatial along-track resolution of ICESat-2. A multiple linear regression model was developed to correct the bias between the two satellites, enabling the retrieval and construction of long-term series WSE for thermokarst lakes of differing sizes with a Root Mean Square Difference (RMSD) of 0.03 m. In the final research chapter, owing to the spatial and temporal limitations in existing altimetry satellites, a proof-of-concept methodology was developed for the use of the newly launched Surface Water and Ocean Topography (SWOT) Ka-band RADAR Interferometer (KaRIn) altimetry satellite in the retrieval of WSE of thermokarst lakes. The chapter's findings revealed that open water and near-land water classes are the most suitable for thermokarst lake WSE retrieval, with a Sigma-1 error of 0.28 m and a Mean Absolute Deviation (MAD) of 0.21 m. The frequency of pixel clouds over a lake and wind speed contribute to the retrieved WSE error. The outcomes of this research laid the foundation for developing a larger framework to monitor long-term water storage and model future Arctic-Boreal hydrological processes.Canadian Space Agency2026-12-1