22484 research outputs found
Sort by
Evaluating the importance of cryptic reservoirs of species and functional groups for biodiversity maintenance in forests
No full textBiodiversity is essential to the functioning and resilience of ecosystems, but some cryptic elements of biodiversity are often overlooked and undervalued. Nonetheless, these elements, which include functional traits and long-term coexistence strategies, are still affected by common processes such as disturbance and herbivory. Yet, we lack the data to reliably predict their responses. In particular, the responses of traits such as the timing of reproduction and the ability to form a viable seed bank, which have far-reaching impacts on the rest of the community, have not been thoroughly explored in the context of chronic herbivory. Furthermore, the consequences of overbrowsing on a seed bank community that has experienced interacting natural and anthropogenic disturbances are unknown. Generally, the importance of forest seed banks has been underestimated. Here, I show that the impact of herbivory on these reproductive traits is highly context dependent, rather than unilaterally negative. I also illustrate the importance of the soil seed bank in forests, both in facilitating community regrowth following intense interacting disturbances, but also in terms of the sheer diversity of species that seed bank in forests worldwide. To date, studies on the conditions of forest seed banks in different biomes have produced highly variable results. Furthermore, my research highlights that studies of seed banking that are integrated with the many interacting processes that shape plant communities are extremely rare and desperately needed. On the whole, this work calls for a greater appreciation of the less-apparent axes of biodiversity that are impacted by foundational ecosystem processes, and an incorporation of those cryptic elements of biodiversity into the narrative on how disturbance and herbivory change forest communities
Integrating EPR and computational modeling to measure protein structure and dynamics
No full textElectron paramagnetic resonance (EPR) spectroscopy coupled with site-directed spin labeling has become a powerful method for probing conformational diversity and dynamics of macromolecules. The sparse distance and dynamics information obtained from EPR measurements greatly benefit from computational modeling. In Chapter 1 of this dissertation, I provide a comprehensive overview of different modeling techniques that can be coupled with EPR. These computational approaches can be used to sample protein and label conformations, simulate EPR spectra, predict or refine protein structures and capture large-amplitude conformational transitions. In Chapter 2, I describe the development of new force field parameters for double histidine- copper(II) (dHis)-Cu(II)-based EPR labels. Molecular dynamics (MD) simulations based on these new force fields generate distance distributions between the labels in remarkable agreement with experiments. These MD-trajectories help us understand the orientational selectivity in double electron-electron resonance (DEER) using Cu(II)-based labels. In Chapter 3, I showcase a new strategy that enables sampling conformational changes at atomic resolution by combining dHis- Cu(II) EPR and weighted ensemble MD simulations. This strategy has been applied to sample a seconds-timescale conformational change in the homodimeric detoxification enzyme. These simulations reveal the negative cooperativity within the enzyme controlled by key residue-residue interactions, which may be essential for the enzyme to protect cells from a broad range of toxins. In Chapter 4, I discuss the development and application of an in silico approach to optimize DEER data acquisition in collaboration with my coworker. This optimal DEER acquisition scheme improves the efficiency of obtaining Cu(II)-based EPR distance distributions, and reduces the data collection time by as much as six fold. Overall, this body of work presents the potential of integrating EPR measurements and computational modeling to tackle various biophysical questions over a wide range of timescales
Exploring ML-Oriented Hardware for Accelerated and Scalable Feature Extraction
No full textMachine-learning (ML) algorithms, tools, and devices continually grow intending to automate and accelerate many aspects of daily life. Hardware accelerators can enable these ML apps to achieve maximized performance. The first phase of this dissertation explores the maximum throughput performance of field-programmable gate arrays (FPGAs), CPUs, and GPUs on two architecturally different convolutional neural networks (CNNs) that are comprised of similar fundamental neural-network operations: GoogLeNet and AlexNet. Because of their highly parallel nature, GPUs achieved the highest inference throughput across models and devices, where additional tensor acceleration significantly boosts performance.
To better understand the design and impacts of ML-oriented hardware and software, the second phase of this dissertation analyzes the subsequent generations of high-performance and embedded devices that feature ML optimizations in terms of latency and throughput. Tensor, vision, and other ML-focused architectures are also considered. Because many of these devices feature hardware for quantized and reduced-precision datatypes, GoogLeNet and AlexNet are quantized with more modern ML frameworks for optimized performance with state-of-the-art backend acceleration software. Though GPUs dominate in throughput and FPGAs achieve the lowest latencies, all of the devices use significant compute, memory, and power resources to achieve their respective performance.
The final phase of this dissertation explores neuromorphic technology as an alternative solution to ML object classification to reduce the overall compute, memory, and power required. Neuromorphic sensors capture events at a microsecond resolution as opposed to generating entire frames to limit the amount of redundant data captured. These events can be related spatially, through algorithms such as k-means clustering, or spatio-temporally, through neuromorphic algorithms such as "A Hierarchy Of event-based Time Surfaces" (HOTS). FPGA accelerators for k-means clustering and HOTS are designed and optimized using state-of-the-art high-level synthesis tools and evaluated on multiple datasets. The highly scalable k-means clustering accelerator achieved an event-processing latency of 65 nanoseconds and throughput of 15.38 MEvt/s while using less than 2% of available FPGA resources and being competitive in accuracy. This dissertation benchmarked many state-of-the-art hardware accelerators, analyzed the impacts of ML hardware and software optimizations, and developed an ultra-low-latency, scalable alternative to ML object classification with neuromorphic technology
Building spaces for play: How mothers design and explore new play environments with pre-walking and walking infants
No full textThe physical environment is the backdrop for infant development. Yet researchers know little about how infants’ spaces come to be. For infants, play spaces are typically structured by caregivers who choose objects, furnishings, and their organization. As infants’ locomotor skills develop so does their agency for engaging with space. Learning to walk, for example, changes how infants interact within the environment. Compared to pre-walkers, walkers move more, travel more, and spend more time playing at a distance from caregivers. As a result, caregivers likely update the arrangements of infants’ spaces in response to an advancing repertoire for action. But how do caregivers build spaces for infant play? Does infant motor ability shape the spaces caregivers construct? This dissertation introduced a novel paradigm to examine how mothers of pre-walking and walking infants created a new play space and engaged in play in different environments. We observed 52 12-month-old infants (35 pre-walkers, 17 walkers) and their mothers. Mothers were asked to design a playroom using a set of building blocks in an empty room. We examined relations among infants’ locomotor status, mothers’ design choices, and patterns of infant and mother behavior during play (eight minutes in an identically organized, standard playroom; eight minutes in mother-designed playrooms). Mothers built spacious playrooms for pre-walkers (multiple constructions spanning a large area) and concentrated playrooms for walkers (one construction covering a small area). In the standard playroom, walkers moved more, traveled more, and engaged in more complex patterns of interpersonal distance to mothers compared to pre-walkers. Mothers directed similar amounts of language and gesture to infants, but communicated more frequently while infants were moving. In mother-designed playrooms, patterns of infant and mother behavior were similar. Differences in mothers’ playroom design (indexed by built area) only related to one behavior: infants generated more room layout changes (by moving blocks) when playrooms were larger. Taken together, this study expands our understanding of how infant motor development shapes caregiver behavior by extending connections to spatial construction. Most importantly, we contribute new insights about the dynamics of infant and caregiver behavior as a process embedded in the physical environment
Design and Processing of Highly Conductive Transparent Electrodes for Semi-transparent Solar Cells
No full textSolar energy, the fastest growing energy source, reduces reliance on fossil fuels and mitigates global warming. A single-junction silicon solar cell is dominant over other types of solar cells, but it has limitations in further increasing theoretical efficiency and diversifying installation locations in residential areas. Multiple semi-transparent thin film solar cells are proposed to address these problems. They can be applied to windows of residential buildings and stacked with the single-junction silicon solar cell to increase the theoretical efficiency above 35%. To advance the development of semi-transparent solar cells, it is necessary to increase both optical transparency and electrical conductivity of transparent electrodes. However, there is an inherent tradeoff between the two properties.
This dissertation aims 1) to explore a property-structure-process relationship in highly conductive transparent electrodes of metal – oxide multilayers using intense pulsed light, and 2) to design unique semi-transparent solar cells with advanced light management and new transparent electrodes.
First, a rational design of a solar cell structure is proposed to overcome the tradeoff between light absorption and optical transparency. Fabry-Perot resonance is incorporated to enhance light absorption in a limited thickness of photo-active layer. By using the wavelength selective property of one-dimensional photonic crystals, device efficiency is increased without sacrificing visible transparency.
Second, metal – oxide multilayer films of high optical transparency and electrical conductivity are studied. Extremely short heat-treatment by intense pulsed light is employed to not only manipulate the dewetting of an ultrathin metal layer but also crystallize both metal and oxide layers. This allows to utilize the high electron concentration of the metal layer and the high carrier mobility of the oxide layer without deteriorating the theoretically predicted optical transparency.
Finally, a near infrared transparent solar cell is designed using the new transparent electrode to build four-terminal tandem solar cells. The tandem device consists of a semi-transparent halide perovskite solar cell and a silicon solar cell. A combination of experimental and theoretical studies clearly exhibits the potential of the new multilayer transparent electrode for the highly efficient tandem solar cells
Cerebral Blood Flow as a Possible Neurobiological Mechanism Explaining the Relationship Between Physical Activity and Depressive Symptoms in Older Adults
No full textThe antidepressant effects of physical activity and exercise in clinical and subclinical populations is well supported, but the neurobiological underpinnings of this relationship are poorly understood. Current evidence suggests that cerebral blood flow (CBF) to the anterior cingulate cortex (ACC) and hippocampus may play a role in depression and responds to chronic exercise. Leveraging baseline cross-sectional data from the Investigating Gains in Neurocognition in an Intervention Trial of Exercise (IGNITE), a multi-site aerobic exercise intervention, the present study aimed to (1) examine the relationship between moderate to vigorous physical activity (MVPA) and sub-clinical depressive symptoms, (2) examine the relationship between MVPA and CBF to the ACC and hippocampus, and (3) test CBF to the ACC and hippocampus as a statistical mediator between MVPA and depressive symptoms in a sample of 544 older adults. MVPA was measured using a wrist-worn accelerometer (Actigraph) and processed using GGIR. CBF was measured using a pseudo-continuous arterial spin labeling sequence for MRI, and regions of interest were extracted using FreeSurfer-derived masks. Aims 1 and 2 were tested using linear regression, and Aim 3 was tested using multiple mediation. All analyses controlled for age, race, gender, years of education, study site, BMI, and past smoking status. Results support a non-linear relationship between MVPA and depressive symptoms such that greater daily levels of MVPA were associated with fewer depressive symptoms, especially for those engaging in close to thirty minutes of MVPA per day. However, we did not find evidence for a relationship between MVPA and CBF to the ACC and hippocampus, nor significant mediation of CBF to the ACC and hippocampus in the relationship between MVPA and depressive symptoms. Our results confirm the relationship between MVPA and depressive symptoms, and do not support a mediating role of CBF in the relationship between MVPA and depressive symptoms
A Minimal Invariant Set, Frames, and Operators
No full textIn 1981, Dale Alspach produced the first example of a non-expansive mapping T on a weakly compact convex subset of a Banach space that is fixed point-free. Using Zorn’s lemma, there exists a minimally weak compact, convex subset of which is T-invariant and fixed point free. We will show that a special closed linear span contains a copy of the space of Lebesgue integrable function on the unit interval. Wenchang Sun introduced g-frames which are generalized frames and include ordinary frames. We will use many ideas from Operator Theory to show a characterization of Frames in terms of Riesz bases and a characterization of g-Frames in terms of g-Riesz bases
Title Page Exploring the Influence of Progranulin on Vascular Contractility and Mitochondrial Homeostasis
No full textCardiovascular diseases (CVDs) encompass a spectrum of conditions impacting the heart and blood vessels, attributing to around 19.1 million global deaths in 2020. These diseases can be congenital, acquired, or have a hereditary basis. Vascular smooth muscle cells (VSMCs) are crucial in maintaining vessel structure, diameter regulation, and vascular calcification—central processes in CVDs. Progranulin (PGRN), a glycoprotein, plays diverse roles in various tissues and cell types, involving embryogenesis, inflammation, wound healing, neurodegeneration, and lysosomal function. Mutations in the granulin (GRN) gene causing PGRN protein insufficiency lead to neurodegeneration. Our research reveals the significance of PGRN in protecting against CVDs.
Analyzing aorta from PGRN-deficient mice, we discovered reduced aortic contractility, unveiling a new facet of PGRN’s function. RNA sequencing showed suppressed oxidative phosphorylation (OXPHOS) in aortae from PGRN knockout mice, notably downregulating genes related to complex I activity in the electron transport chain pathways.
PGRN's functions span cell growth, embryogenesis, anti-inflammatory responses, and wound healing. We propose studying the molecular mechanisms behind mutant PGRN, potentially driving cardiovascular defects. Primary VSMCs from PGRN knockout mice exhibited perturbed contraction, reduced oxygen consumption rate (OCR), and impaired mitophagy, paralleled by lower ATP levels and increased mitochondrial oxidative stress due to PGRN deficiency.
Increasing PGRN expression through a viral vector significantly improved vascular contractility, OCR, mitochondrial complex I activity, and reduced oxidative stress. PGRN deficiency disrupted lysosomal function within VSMCs, affecting mitophagy. In a model with chronic Angiotensin II treatment, PGRN-deficient mice displayed unresponsiveness in hypercontractility and increased collagen deposition.
Our research underscores PGRN's crucial role in maintaining vascular contractility by regulating mitochondrial function. With the need for novel CVD therapies, these findings are significant. PGRN's influence on mitochondrial complex I quality, mitochondria recycling, and redox signaling pathways suggest its potential as a therapeutic target for CVDs, positioning PGRN as a promising alternative in treatment strategies
Characterization of phenotypes resulting from the ectopic expression of T and Tbx6 in the developing limb of Mus musculus and Drosophila melanogaster
No full textMembers of the evolutionarily conserved T-box transcription factor family share a similar DNA binding domain (T-domain) and thus have the potential to recognize similar DNA sequences in target genes. Outside of this T-domain the proteins are diverse suggesting that they function differently to regulate transcription. Interestingly, several human T-box genes show haploinsufficiency, including Tbx5 and Tbx3 that when mutated cause Holt-Oram and ulnar- mammary syndrome, respectively. This dosage sensitivity combined with the shared T-domain lead to our hypothesis that when co-expressed a competition between T-box factors at target genes can occur. I used the mouse and Drosophila limb systems to investigate the effect of artificially changing the relative levels of T-box proteins expressed in the developing limbs by ectopically expressing T or Tbx6. I characterized the phenotypes in both model systems. In the mouse, ectopic T and Tbx6 resulted in a smaller limb and skeletal defects reminiscent of hypomorphic Tbx5 limbs, and yet the Tbx6 limbs exhibited a digit duplication that is not present in T limbs. The Drosophila wing is sensitive to the levels of the Tbx2 ortholog omb. Using the GAL4/UAS system we have ectopically expressed either mouse T or Tbx6 in the developing wing disc and observed different adult wing phenotypes depending on the T-box expressed. To further examine the potential mechanism causing these different phenotypes and to understand the relationship between T-box family members, I have initiated studies to ectopically express both activating and repressive versions of T and Tbx6
Collaborative Goal Setting in the Family Check-Up: Do parents' goals relate to improvement in positive behavior support and child behavior?
No full textGoal setting is used in many parenting interventions to tailor treatment to families’ needs and bolster motivation. The present study describes the thematic content of parents’ treatment goals in the early childhood Family Check-Up (FCU) and investigates whether goals focused on positive behavior support (PBS) or child externalizing behavior at child age 2 relate to improvements in these domains one year later and examines moderators of such associations. Participants included 250 families who set goals at child age 2 in the intervention arm of the Early Steps Multisite Study, which is a randomized controlled trial of the FCU among 731 families who were screened for risk and recruited from Women, Infants, and Children (WIC) clinics. Family assessments occurred at child ages 2 and 3 and included observational assessment of parenting. Setting a PBS goal and/or a child externalizing behavior goal were not related to growth in observed PBS or child externalizing behavior, respectively, nor did parental self-efficacy, parental social support, or attendance in follow-up treatment sessions moderate these relationships. These results suggest that setting goals to improve PBS and child externalizing behavior may not uniquely contribute to established treatment effects of the FCU in these domains. Findings run contrary to theory suggesting that goals motivate behavior change. If replicated, null findings may inform adaptations to the use of and/or methods for goal setting in the FCU