Understanding the performance of applications on modern multi- and manycore platforms is a difficult task and involves complex measurement, analysis, and modeling. The Roofline model is used to assess an application's performance on a given architecture. Not much work has been done with the Roofline model using real measurements. Because it can be a very useful tool for understanding application performance on a given architecture, in this thesis we demonstrate the use of architectural roofline data with measured data for analyzing the performance of different benchmarks. We first explain how to use different toolkits to measure the performance of a program. Next, these data are used to generate the roofline plots, based on which we can decide how can we make the application more efficient and remove bottlenecks. Our results show that this can be a powerful tool for analyzing performance of applications over different architectures and different code versions.

The desire for high performance on scalable parallel systems is increasing the complexity and the need to tune MPI implementations. The MPI Tools Information Interface (MPI T) introduced in the MPI 3.0 standard provides an opportunity for performance tools and external software to introspect and understand MPI runtime behavior at a deeper level to detect scalability issues. The interface also provides a mechanism to fine-tune the performance of the MPI library dynamically at runtime. This thesis describes the motivation, design, and challenges involved in developing an MPI performance engineering infrastructure using MPI T for two performance toolkits — the TAU Performance System, and Caliper. I validate the design of the infrastructure for TAU by developing optimizations for production and synthetic applications. I show that the MPI T runtime introspection mechanism in Caliper enables a meaningful analysis of performance data. This thesis includes previously published co-authored material.

Football safety has increased over time, in part due to improvements in equipment and body mechanics, but there are still inherent risks involved, including exposure to repetitive head impacts (RHI). Significant head impacts can result in a constellation of symptoms including nausea, vomiting, headache, dizziness, and amnesia, which typically assist in the diagnosis of concussion. However, it has been shown that subconcussive impacts may result in microstructural changes and physiological alterations in the brain. This is particularly concerning because athletes may be undergoing changes in the brain in the absence of outwardly visible symptoms. Poorer neurologic outcomes later in life have been associated with cumulative exposure rather than number of diagnosed concussions. Accelerometers installed in helmets have shown that college football players may receive up to 1,850 head impacts throughout the course of one season. The concussion rate is obviously much lower, indicating there are a high number of head impacts per diagnosed concussion. Axons are especially susceptible to damage from RHI because of their extension throughout the nervous system. The subtle changes thought to result from RHI are not easy to measure, but several modalities have been proposed. These include diffusion tensor imaging (DTI), plasma tau protein, and King-Devick testing. The proposed study will look to quantify cumulative head impact exposure in college football players prior to the start of a season and see if this has any impact on the variables. They will then participate in one season of football wearing helmet accelerometers to measure the number of head impacts sustained. Changes in the variables will be compared to non-contact sport college athletes. Data will be analyzed to determine if number of head impacts correlates with changes in variables and if prior head impact exposure has any effect on these changes. Data obtained from this study will have significant implications in the field of head injury. It may strengthen the use of several markers of brain injury that could be utilized in the future. Additionally, the effects of cumulative head impact exposure and one season of head impacts will be thoroughly examined. This information can be provided to trainers, coaches, and athletes to further improve football safety.

RNA binding proteins (RBPs) have been found to be frequently involved in neurodegenerative diseases (Ash 2014). Mutations in RBPs cause amyotrophic lateral sclerosis (ALS), spinocerebellar ataxia, frontotemporal dementia (FTD) and myopathies (Ash 2014), and recent studies suggest that aggregation of RBPs are a pathological feature frequently encountered in tauopathies (Vanderweyde 2016). Emerging studies on neurodegenerative diseases are now showing an increasingly important role for tau in regulating the biology of RBPs. In this study, we examine findings that show that reducing the RBP T-cell intracellular antigen 1 (TIA1) in vivo not only protects against neurodegeneration, but also prolongs the survival rate in transgenic P301S tau mice. Furthermore, the reduction of TIA1 decreases the number and size of granules co-localizing with stress granule markers, and inhibits the accumulation of tau oligomers, although at the expense of an increased number of neurofibrillary tangles. However, despite the observed increase in neurofibrillary tangles, this TIA1 reduction still manages to increase neuronal survival, rescue behavioral deficits and prolong lifespan. The in vivo data presented in this study suggests an important role for TIA1 in mediating toxicity and provides evidence that RBPs orchestrate a pathway to tau aggregation and the resulting neurodegeneration.

Ce travail est issu de plusieurs constats. Premièrement, il est admis aujourd’hui que la capacité à percevoir et à retirer les informations visuelles les plus pertinentes d’une action adverse est un facteur déterminant de la performance sportive et ce d'autant plus si l'action est incertaine et faussée. Le second constat, découlant du premier, est que la tromperie est mise en jeu dans bon nombre d'interactions humaines, mais qu'elle est particulièrement déterminante et prépondérante dans lecadre des stratégies sportives. Enfin, le dernier constat est que l'exploration de la performance sportive, pour des raisons méthodologiques ou théoriques, ne permet pas dans certains cas d'expliquer pourquoi un expert est meilleur qu'un novice.L’objectif de notre travail est de comprendre ces déterminants (perceptivo-moteurs) de l'expertise lors d’un duel 1 vs. 1 en rugby présentant une tentative de feinte de l’attaquant. Nous résumons ce travail en trois questions concrètes auxquelles nous répondons: Comment réaliser une feinte de corps en rugby? Comment la détecter? Comment y faire face? La première étude s'attache à analyser les différences biomécaniques observables entre un mouvement de simple changement de direction et un mouvement de feinte de corps en rugby. La seconde étude, en immersion virtuelle, met en lien les informations prospectives émanant des mouvements de feinte et la prise de décision du défenseur de rugby (expert ou novice). Enfin, en se basant sur une méthodologie quasi similaire, la troisième étude vient immerger les participants dans une situation complète de défense, nous permettant d'explorer leur stratégie perceptivo-motrice. L'ensemble de ce processus nous permet de témoigner de l'effet de l'expertise mais surtout d'expliquer pourquoi l'expert est meilleur en dressant les différences de stratégie perceptivo-motrice le caractérisant. Si les implications concernent majoritairement le domaine du rugby (connaissances fondamentales et entrainement), elles présentent au sens plus large, un éclairage justifié et démontré des processus mis en jeu lors de toute interaction, humaine ou animal, sollicitant la tromperie. Nous en sommes persuadés; un prédateur animal, un joueur de poker, un défenseur de basket... se doivent deprésenter des stratégies quasi-similaires pour détecter l'intention de bluff chez l'opposant / This work comes from several observations. First, it is now accepted that the ability to pick up and attend to the most relevant visual information from an opponent action is a determinant of performance and even more so if the action is uncertain and deceptive. The second observation, deriving from the first, is that deception is put into play in many human interactions, and it is particularly crucial in sports strategies. Finally the last observation is that exploration of athletic performance, for methodological or theoretical reasons, can’t explain, in certain cases, why an expert is better than a novice.The aim of our work is to understand these determinants (perception and action) of expertise in a 1 vs. 1 duel in rugby, presenting an attempt of deceptive movement by the attacker. We summarize this work in three specific questions we answer: How to make a deceptive movement? How to detect a deceptive movement? How to answer to a deceptive movement? The first study aims to analyse the biomechanical differences between a movement with a simple changeof direction and a deceptive movement in rugby. The second study, using virtual reality, links the prospective information of the attacker’s movement and the decision making of the rugby defender (expert or novice). Finally, based on a quasi-similar methodology, the third study immerses the participants in a situation requiring a complete action of defence, allowing us to explore the perception and action strategies of the participants. This whole process allows us to demonstrate the effectof expertise but also to explain why experts are better by drawing up the differences characterizing them in terms of perception and action strategies. If the implications mainly concern the field of rugby (basic knowledge and training), they are also of great interest for any interaction, human or animal, involving deception. We are convinced, a predatory animal, apoker player, a defender of basketball ... must present similar strategies to detect the deceptive intention of an opponent

Orientador: Daniel Carneiro Carrettiero. / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Biossistemas, 2012.

abstract: Alzheimer’s Disease (AD) and Frontotemporal Dementia (FTD) are the leading causes of early onset dementia. There are currently no ways to slow down progression, to prevent or cure AD and FTD. Both AD and FTD share a lot of the symptoms and pathology. Initial symptoms such as confusion, memory loss, mood swings and behavioral changes are common in both these dementia subtypes. Neurofibrillary tau tangles and intraneuronal aggregates of TAR DNA Binding Protein 43 (TDP-43) are also observed in both AD and FTD. Hence, FTD cases are often misdiagnosed as AD due to a lack of accurate diagnostics. Prior to the formation of tau tangles and TDP-43 aggregates, tau and TDP-43 exist as intermediate protein variants which correlate with cognitive decline and progression of these neurodegenerative diseases. Effective diagnostic and therapeutic agents would selectively recognize these toxic, disease-specific variants. Antibodies or antibody fragments such as single chain antibody variable domain fragments (scFvs), with their diverse binding capabilities, can aid in developing reagents that can selectively bind these protein variants. A combination of phage display library and Atomic Force Microscopy (AFM)-based panning was employed to identify antibody fragments against immunoprecipitated tau and immunoprecipitated TDP-43 from human postmortem AD and FTD brain tissue respectively. Five anti-TDP scFvs and five anti-tau scFvs were selected for characterization by Enzyme Linked Immunosorbent Assays (ELISAs) and Immunohistochemistry (IHC). The panel of scFvs, together, were able to identify distinct protein variants present in AD but not in FTD, and vice versa. Generating protein variant profiles for individuals, using the panel of scFvs, aids in developing targeted diagnostic and therapeutic plans, gearing towards personalized medicine. / Dissertation/Thesis / Doctoral Dissertation Neuroscience 2018

The APOE gene is of profound importance regarding the onset of Alzheimer's disease (AD). From the small physical differences among the protein products of the isoforms of this gene arises a profound difference in their physiologies. For example, the APOE2 isoform confers resistance to AD, the APOE3 isoform confers neutral susceptibility to AD, and the APOE4 isoform confers proneness to AD. L-cysteine is an amino acid that has several anti-AD properties, among which are its ability to sequester iron and form glutathione – a powerful antioxidant – and therefore may be a promising potential dietary supplement for ameliorating AD pathology. In our experiment, we fed Mus musculus (mice) homozygous for APOE2, APOE3, and APOE4 either a control diet or a diet high in L-cysteine. Using Western blotting analysis, we quantified Amyloid β (Aβ), hyper-phosphorylated Tau (HP-Tau), and the three APOE proteins that we extracted from post-mortem brains of APOE2, APOE3, and APOE4 homozygous mice of 3-, 6-, 9-, and 12-month ages. We calculated a three-way ANOVA on a sample of 86 mice to examine the effect of age, genotype, and diet on protein quantities. We found that administration of L-cysteine trends towards lowering levels of Aβ in each cohort, but this effect is statistically insignificant. On the other hand, L-cysteine caused a significant decrease in APOE production with regard to diet [F(1,62) = 6.17, p=0.02], indicating that less APOE is produced due to the decrease in Aβ burden. Furthermore, administration of L-cysteine revealed no significant impact on or trends regarding HP-Tau deposition between diet types for each cohort. However, we observed that L-cysteine appeared to nullify the increasing trend in HP-Tau deposition between APOE2 and APOE4 cohorts. Thus, L-cysteine may be weakly affecting HP-Tau deposition via its ability to somewhat reduce Aβ burden and consequently prevent the shutdown of the proteosomes responsible for the degradation and clearance of HP-Tau. Taken together, these data suggest that L-cysteine should be considered as an intervention for AD pathology.

Tauopathies are a group of neurodegenerative diseases characterized by the abnormal deposition of the protein tau, a microtubule stabilizing protein. Under normal physiological conditions tau is a highly soluble protein that is not prone to aggregation. In disease states alterations to tau lead to enhanced fibril formation and aggregation, eventually forming neurofibrillary tangles (NFTs). The exact cause for NFT deposition is unknown, but increased post-translational modifications and mutations to the tau gene can increase tangle formation. Tauopathic brains are stuck in a detrimental cycle, with cellular dysfunction contributing to the development of tau pathology and the development of tau pathology contributing to cellular dysfunction. The exact mechanisms by which each part of the cycle contributes to the other are still being explored. To investigate the unique contributions of each part of this cycle we utilized two separate models of tauopathy: one chronic and one acute. Overall this project provides novel insight into the role of pathological tau as both a cause, and a consequence, of cellular dysfunction. To understand how development of tau pathology contributes to cellular dysfunction we studied chronic disease models. Using human brain tissue we found that under normal conditions tau associates with ribosomes but that this interaction is enhanced in Alzheimer’s disease brains. We then used in vitro and in vivo models of tauopathy to show that this association causes a decrease in protein synthesis. Finally, we show that wild type tau and mutant tau reduce protein translation to similar levels. To understand how general cellular dysfunction contributes to development of pathology we used an acute model of tauopathy through traumatic brain injury (TBI). We injured rTg4510 tau transgenic mice at different ages to investigate the effect of TBI on tau fibrillization (2 month old) and the effect of TBI on tau already in NFTs (4.5 month old). In 2 month old mice, we found that tau hyperphosphorylation was decreased at 24 hours and increased at 7 days post injury, and that tau oligomerization was decreased at 24 hours post injury. We also found that tau fibrillization was not increased after 24 hours or 7 days post injury. In 4.5 month old mice, we found that TBI did not increase or decrease tangle counts in the brain, but we did qualitatively observe decreased variability within groups. Overall these studies contribute novel understanding of tau’s role in different disease states. We identified a functional consequence of the interaction between tau and ribosomes, and demonstrated that a single head impact did not increase tau fibril formation within 7 days of injury. While human diseases associated with TBI show neurofibrillary tangle deposition, we have yet to recreate that aspect of the disease in research models of TBI. Our findings support the need for further investigation into the nuances of tau in disease, especially following TBI.

Neurons are polarized cells with dendrites and an axon projecting from their cell body. Due to this polarized structure a major challenge for neurons is the transport of material to and from the cell body. The transport that occurs between the cell body and axons is called Axonal transport. Axonal transport has three major components: molecular motors which act as vehicles, microtubules which serve as tracks on which these motors move and microtubule associated proteins which regulate the transport of material. Axonal transport maintains the integrity of a neuron and its dysfunction is linked to neurodegenerative diseases such as, Alzheimer’s disease, Frontotemporal dementia linked to chromosome 17 and Pick’s disease. Therefore, understanding the process of axonal transport is extremely important. Single particle tracking is one method in which axonal transport is studied. This involves fluorescent labelling of molecular motors and microtubule associated proteins and tracking their position in time. Single particle tracking has shown that both, molecular motors and microtubule associated proteins exhibit motion with multiple components. These components are directed, where motion is in one direction, diffusive, where motion is random, and static, where there is no motion. Moreover, molecular motors and microtubule associated proteins also switch between these different components in a single instance of motion. We have developed a MATLAB program, called MixMAs, which specializes in analyzing the data provided by single particle tracking. MixMAs uses a sliding window approach to analyze trajectories of motion. It is capable of distinguishing between different components of motion that are exhibited by molecular motors and microtubule associated proteins. It also identifies transitions that take place between different components of motion. Most importantly, it is not limited by the number of transitions and the number of components present in a single trajectory. The analysis results provided by MixMAs include all the necessary parameters required for a complete characterization of a particle’s motion. These parameters are the number of different transitions that take place between different components of motion, the dwell times of different components of motion, velocity for directed component of motion and diffusion coefficient for diffusive component of motion. We have validated the working of MixMAs by simulating motion of particles which show all three components of motion with all the possible transitions that can take place between them. The simulations are performed for different values of error in localizing the position of a particle. The simulations confirm that MixMAs accurately calculates parameters of motion for a range of localization errors. Finally, we show an application of MixMAs on experimentally obtained single particle data of Kinesin-3 motor.