Although adult neurons in the mammalian central nervous system do not spontaneously regenerate after injury, growing evidence indicates that genetically manipulating them can increase their ability to regenerate their axons. However, as genetic manipulation is not clinically feasible, current research continues to investigate pharmacological approaches to transiently enhance the intrinsic ability of adult CNS neurons to survive and regenerate their axons. Amphiregulin is a unique epidermal growth factor receptor ligand that has been shown to be crucial in liver regeneration, and accumulating evidence suggests that AREG signaling can promote both survival and axon regeneration of neurons. Growth differentiation factor 11 is a member of the transforming growth factor β superfamily, and it has been shown to exert rejuvenation effects in the aged brain and promote neuronal survival in the CNS. However, it is unknown whether AREG and GDF11 can induce neuronal survival and axon regeneration in the visual system. Hence, the current study investigated: 1) the developmental and post-injury expression pattern of these two ligands and their respective receptors, EGFR and activin-like kinase 5, 2) the potential neuroprotective and regenerative effects of these two ligands on retinal ganglion cells by using optic nerve crush model, and 3) the molecular mechanisms mediating the neuroprotective and regenerative effects of AREG and GDF11 on RGCs. Based on the western blot and immunohistochemical data, both AREG and GDF11 are consistently expressed throughout the retina development, but their receptors are only upregulated during the early retina development. Furthermore, AREG expression is significantly reduced in the adult retina 7 days after ONC whereas GDF11 expression remains unchanged after ONC. Interestingly, while EGFR expression is only upregulated 3 days after ONC, ALK5 expression is consistently upregulated throughout the post-injury time course. A single intravitreal injection of AREG or GDF11 immediately after ONC promoted significant RGC survival by activating Smad2/3 pathway. Because both ligands were not able to promote RGC axon regeneration, activation of Smad2/3 pathway may promote RGC survival but suppress RGC axon regeneration. Overall, the findings indicate that both AREG and GDF11 hold therapeutic potential for both neurodegenerative diseases and retinal degenerative diseases.
The COVID-19 lockdowns had negative impacts on psychological well-being and amplified certain stressors (e.g., social isolation), especially in at-risk populations. My thesis examined social support availability as a coping strategy. Single people living alone (N=220) were recruited during the initial 2020 lockdown and followed over a period of six weeks. Each week, participants reported their perceived social support availability, social isolation, and life satisfaction. I hypothesized that greater perceptions of social support availability, both on the within- and between-person levels, would buffer the negative effects of social isolation on psychological well-being. Multi-level modeling results showed that stress-buffering occurred on the between-person level. Associations were analyzed longitudinally, revealing that lagged social isolation did not predict life satisfaction the week after. An interaction was observed between lagged social isolation and lagged social support availability, such that lagged social isolation predicted less life satisfaction when social support was unavailable the week before.
The ATLAS Experiment measures the properties of particles created in the high-energy proton-proton collisions delivered by the Large Hadron Collider at CERN. The Standard Model of particle physics is our best description of the subatomic world - this well-studied theory provides accurate and precise descriptions of the fundamental particle properties and their interactions. Multiple areas of the Standard Model make predictions that are more precise than the corresponding experimental measurements - in some cases, increased experimental precision could suggest significant discrepancies between the Standard Model prediction and experimental measurement, potentially hinting at new physics to elucidate. One such area of tension between prediction and experiment is in the weak sector, whose force is mediated by W and Z bosons. The current best predictions based on the Standard Model do an inadequate job of precisely predicting one important observable of W and Z bosons: their transverse momentum, p_T. Precision measurements of properties like the p_T of W and Z bosons improve our knowledge of the weak sector, and are vital stepping stones to critical measurements like the mass of the W boson, whose most recent reported measurement is inconsistent with the Standard Model prediction. In this thesis, I explain my work using ATLAS data to make high-precision measurements of the p_T of W and Z bosons using the decay channels W --> l nu and Z --> l l (l = e, mu) at centre-of-mass energies 13 and 5 TeV using a special low-pileup dataset. I show that I have helped reduce the systematic uncertainties to percent-level precision and that statistical uncertainties are dominant, which demonstrates that more low-pileup data should be taken in order to further reduce the total uncertainty to eventually help resolve Standard Model weak-sector discrepancies like that of the W boson mass. I also detail my work to improve the way that electrons are identified by the ATLAS detector using a technique called W tag-and-probe. In particular, I validated the use of a new trigger designed specifically for electron identification with the W tag-and-probe technique.
Quantum field theory (QFT) is a conceptual framework for understanding the behaviour of subatomic particles—the most successful, mathematically rigorous formulation of QFT is in the language of operator algebras. In this thesis, we describe the construction of specific kinds of QFTs using operator-algebraic methods. Once we have described their construction in detail, we use tensor network methods (which are at the centre of modern quantum physics) to build approximations of these QFTs. We finish with a discussion on the relationship between our tensor networks and those used in toy models of the AdS/CFT correspondence.
Biodosimetry relies on calibration curves to convert biological damage induced by ionizing radiation to an absorbed dose. Health Canada generates these curves by irradiating biological samples with X-rays, though exposures scenarios could consist of other types of radiation which are challenging to replicate in the laboratory. The ultimate goal of this work is to model the X-ray setup using Monte Carlo methods and to validate the model using in-laboratory measurements. The model was iterated through preliminary and final testing phases in different EGSnrc applications (egs++, SpekPy and BEAMnrc) and optimized using variance reduction techniques, resulting in multiple models. The X-ray spectra produced from each model were compared and found to be equivalent. Model outputs were also compared against laboratory measurements to identify the most accurate model. The final model output will be used in the next phase of the project to model radiobiological damage.
Social-cognitive impairments in schizophrenia are markers that precede the illness and are present in first-degree relatives of patients. Adolescence and young adulthood are peak ages of risk for the onset of schizophrenia and are important windows to observe impairments that could signify the transition to schizophrenia. To explore these risk markers, this case series study described the social-cognitive profiles of young adults at familial high-risk (FHR) and investigated their relation to symptoms of schizophrenia and schizotypy. In this study, 13 controls and 4 participants at FHR completed assessments measuring symptoms, schizotypy, emotion regulation and recognition, theory of mind, and attributional style. Participants at FHR recognized fewer sad faces than controls but did not show other impairments. Furthermore, greater symptoms and schizotypy were associated with worse performance on some social-cognitive domains. Further investigation with larger samples is needed to explore if difficulty recognizing negative emotions is a risk marker for schizophrenia.
In 1996, Parliament passed s.718.2(e) of the criminal code of Canada. In doing so, recognized the injustices experienced by Indigenous populations at the hands of our criminal justice system. Nearly 26 years later, and Indigenous overrepresentation in custody across Canada is still rising. This pattern of injustice raises the question, what is stopping Indigenous persons from accessing justice? By exploring the field of access to justice and defining access to justice as containing procedural, substantive and symbolic elements, this thesis applied a unique approach to measuring access to justice based on the leading approaches of access to justice research. By focusing on the experiences of defence counsel who work within the Gladue framework of sentencing and applying an expansive conception of access to justice to guide the inquiry, this thesis attempts to shed light on the barriers to accessing justice that are faced by Indigenous persons being sentenced.
Short-term load forecast at residential house level plays a critical role in home energy management system. While a variety of machine learning based load forecasting methods have been proposed, their prediction performance have not been assessed against cyber threats on smart meters which have been increasingly reported. This thesis investigates the vulnerability of four extensively used machine learning algorithms for residential short-term load forecast against cyberattacks, including nonlinear auto regression with external input (NARX) neural network, support vector machine (SVM), decision tree (DT), and long-short-term memory (LSTM) deep learning. The predication performances on these two datasets are compared by using NARX, SVM, DT, and LSTM. Four cyberattack models are investigated, including pulse attack, scale attack, ramp attack, and random attack.
This research addresses the assessment of Higher-Order Thinking Skills (HOTS), such as metacognition, reflection, and problem-solving, in Virtual Learning Environments (VLEs). We particularly focus on the use of process metrics, their combinations, and various analysis methods that allow VLE platforms to perform automated HOTS assessments. Traditional learning assessments rely mostly on outputs and are not suitable for HOTS assessment that requires process observation. As a result, it is a challenge for learners and educators to identify the areas of weakness for customized help when it comes to HOTS. Our objective is to understand the requirements of a VLE-based HOTS assessment framework and explore what process metrics can be used and how they can be analyzed to offer insight into learners' HOTS development. To achieve the above objective, a series of four studies were performed within this research. Study 1 was an initial exploratory investigation that suggested 3D VLEs as a possible HOTS fostering platform though associating their unique affordances to the requirements of common learning theories. This study was a motivational activity and initiated our research. Study 2 was performed on a text-based VLE and provided new insight into how aggregated process metrics can be used to represent student attention and participation, which are linked to HOTS. Study 3 focused on identifying 3D VLE process metrics and their alignment with HOTS components. Study 3 results suggested that the rich data coming from a 3D VLE, and the combination of process metrics as small groups (motifs) and time series, can offer more insights about HOTS. Finally, Study 4 employed motifs and time series-based similarity analysis on process metrics for performing HOTS assessment during learning tasks in a 3D VLE. Study 4 investigated task compatibility with four different similarity indexes, and findings suggested employing different similarity indexes depending on the learning tasks. Overall, the studies conducted within the scope of this research provided supporting evidence of the possibility of automated HOTS assessment on VLEs using process metrics. They showed the value of motifs (small yet meaningful series of process metrics) as a measure for HOTS. However, they suggested that there is no single method, and different learning tasks might use different data analysis strategies.
A major regulatory influence over cell biology is lysine methylation and demethylation within histone proteins. The KDM5/JARID1 sub-family are 2-oxoglutarate and Fe(II)-dependent lysine-specific histone demethylases that are characterized by their Jumonji catalytic domains. This enzyme family is known to facilitate the removal of tri-/di-methyl modifications from lysine 4 of histone H3 (i.e., H3-K4me3/2), a mark associated with active gene expression. As a result, studies to date have revolved around KDM5's influence on disease through their ability to regulate H3-K4me2/3. Recently, evidence has demonstrated that KDM5's may influence disease beyond H3-K4 demethylation, making it critical to further investigate KDM5 demethylation of non-histone proteins. In efforts to help identify potential non-histone substrates for the KDM5 family, we developed a library of 180 permutated peptide substrates (PPS), with sequences that are systematically altered from the WT H3-K4me3 sequence. From this library, we characterized recombinant KDM5A/B/C/D substrate preference. Subsequently we developed recognition motifs for each KDM5 demethylase and used them to predict potential substrates for KDM5A/B/C/D. Demethylation activity was then profiled to generate a list of high/medium/low-ranking substrates for further in vitro validation for each of KDM5A/B/C/D. Through this approach, we analyzed prediction success rate and identified 66 high-ranking substrates in which KDM5 demethylases displayed significant in vitro activity towards. We further shown the ability to monitor changes in cellular methylation in a handful of the 66 high ranking candidate substrates in response to KDM5 inhibition. Specifically, we focused validation efforts on a high-ranking KDM5A novel substrate: p53-K370me3. We demonstrated significant recombinant KDM5A(1-588ΔAP) and KDM5A(1-801) activity towards the p53-K370me3 substrate in vitro. We then monitored KDM5A-mediated demethylation of the p53-K370me3/2 substrate in HCT 116 cells using a combination of wild-type KDM5A and inactive-mutant KDM5A(H483A) overexpression plasmids, along with immunoblotting, (co-) immunoprecipitation and mass spectrometry analysis. Furthermore, we have shown that KDM5A expression influences the established p53-53BP1 interaction. Finally, we identified a novel p53-TAF5 interaction dominated through the p53-K370me3 state and how KDM5A expression might affect this interaction. Ultimately, we have provided the first evidence of a KDM5 demethylase targeting a non-histone substrate for demethylation, via the novel KDM5A demethylation of the p53-K370me3 substrate.