Sparse Signal Representation in Digital and Biological Systems

Theories of sparse signal representation, wherein a signal is decomposed as the sum of a small number of constituent elements, play increasing roles in both mathematical signal processing and neuroscience. This happens despite the differences between signal models in the two domains. After reviewing preliminary material on sparse signal models, I use work on compressed sensing for the electron tomography of biological structures as a target for exploring the efficacy of sparse signal reconstruction in a challenging application domain. My research in this area addresses a topic of keen interest to the biological microscopy community, and has resulted in the development of tomographic reconstruction software which is competitive with the state of the art in its field. Moving from the linear signal domain into the nonlinear dynamics of neural encoding, I explain the sparse coding hypothesis in neuroscience and its relationship with olfaction in locusts. I implement a numerical ODE model of the activity of neural populations responsible for sparse odor coding in locusts as part of a project involving offset spiking in the Kenyon cells. I also explain the validation procedures we have devised to help assess the model's similarity to the biology. The thesis concludes with the development of a new, simplified model of locust olfactory network activity, which seeks with some success to explain statistical properties of the sparse coding processes carried out in the network.

The Role of Attachment in a Social Cognitive Model of Social Domain Satisfaction in College Students

The study examined a modified social cognitive model of domain satisfaction (Lent, 2004). In addition to social cognitive variables and trait positive affect, the model included two aspects of adult attachment, attachment anxiety and avoidance. The study extended recent research on well-being and satisfaction in academic, work, and social domains. The adjusted model was tested in a sample of 454 college students, in order to determine the role of adult attachment variables in explaining social satisfaction, above and beyond the direct and indirect effects of trait positive affect. Confirmatory factor analysis found support for 8 correlated factors in the modified model: social domain satisfaction, positive affect, attachment avoidance, attachment anxiety, social support, social self-efficacy, social outcome expectations, and social goal progress. Three alternative structural models were tested to account for the ways in which attachment anxiety and attachment avoidance might relate to social satisfaction. Results of model testing provided support for a model in which attachment avoidance produced only an indirect path to social satisfaction via self-efficacy and social support. Positive affect, avoidance, social support, social self-efficacy, and goal progress each produced significant direct or indirect paths to social domain satisfaction, though attachment anxiety and social outcome expectations did not contribute to the predictive model. Implications of the findings regarding the modified social cognitive model of social domain satisfaction were discussed.

Exploring the Function and Regulation of Arabidopsis EIN2 in Ethylene Signaling

Ethylene is an essential plant hormone involved in nearly all stages of plant growth and development. EIN2 (ETHYLENE INSENSITIVE2) is a master positive regulator in the ethylene signaling pathway, consisting of an N-terminal domain and a C-terminal domain. The EIN2 N-terminal domain localizes to the endoplasmic reticulum (ER) membrane and shows sequence similarity to Nramp metal ion transporters. The cytosolic C-terminal domain is unique to plants and signals downstream. There have been several major gaps in our knowledge of EIN2 function. It was unknown how the ethylene signal gets relayed from the known upstream component CTR1 (CONSTITUTIVE RESPONSE1) a Ser/Thr kinase at the ER, to EIN2. How the ethylene signal was transduced from EIN2 to the next downstream component transcription factor EIN3 (ETHYLENE INSENSITIVE3) in the nucleus was also unknown. The N-terminal domain of EIN2 shows homology to Nramp metal ion transporters and whether EIN2 can also function as a metal transporter has been a question plaguing the ethylene field for almost two decades. Here, EIN2 was found to interact with the CTR1 protein kinase, leading to the discovery that CTR1 phosphorylates the C-terminal domain of EIN2 in Arabidopsis thaliana. Using tags at the termini of EIN2, it was deduced that in the presence of ethylene, the EIN2 C-terminal domain is cleaved and translocates into the nucleus, where it could somehow activate downstream ethylene responses. The EIN2 C-terminal domain interacts with nuclear proteins, RTE3 and EER5, which are components of the TREX-2 mRNA export complex, although the role of these interactions remains unclear. The EIN2 N-terminal domain was found to be capable of divalent metal transport when expressed in E. coli and S. cerevisiae leading to the hypothesis that metal transport plays a role in ethylene signaling. This hypothesis was tested using a novel missense allele, ein2 G36E, substituting a highly conserved residue that is required for metal transport in Nramp proteins. This G36E substitution did not disrupt metal ion transport of EIN2, but the ethylene insensitive phenotype of this mutant indicates that the EIN2 N-terminal domain is important for positively regulating the C-terminal domain. The defect of the ein2 G36E mutant does not prevent proper expression or subcellular localization, but might affect protein modifications. The ein2 G36E allele is partially dominant, mostly likely displaying haploinsufficiency. Overexpression of the EIN2 N-terminal domain in the ein2 G36E mutant did not rescue ethylene insensitivity, suggesting the N-terminal domain functions in cis to regulate the C-terminal domain. These findings advance our knowledge of EIN2, which is critical to understanding ethylene signaling.

Coarse Graining to Invesitigate Peptide-Lipid Interactions

Experimental characterization of molecular details is challenging, and although single molecule experiments have gained prominence, oligomer characterization remains largely unexplored. The ability to monitor the time evolution of individual molecules while they self assemble is essential in providing mechanistic insights about biological events. Molecular dynamics (MD) simulations can fill the gap in knowledge between single molecule experiments and ensemble studies like NMR, and are increasingly used to gain a better understanding of microscopic properties. Coarse-grained (CG) models aid in both exploring longer length and time scale molecular phenomena, and narrowing down the key interactions responsible for significant system characteristics. Over the past decade, CG techniques have made a significant impact in understanding physicochemical processes. However, the realm of peptide-lipid interfacial interactions, primarily binding, partitioning and folding of amphipathic peptides, remains largely unexplored compared to peptide folding in solution. The main drawback of existing CG models is the inability to capture environmentally sensitive changes in dipolar interactions, which are indigenous to protein folding, and lipid dynamics. We have used the Drude oscillator approach to incorporate structural polarization and dipolar interactions in CG beads to develop a minimalistic peptide model, WEPPROM (Water Explicit Polarizable PROtein Model), and a lipid model WEPMEM (Water Explicit Polarizable MEmbrane Model). The addition of backbone dipolar interactions in a CG model for peptides enabled us to achieve alpha-beta secondary structure content de novo, without any added bias. As a prelude to studying amphipathic peptide-lipid membrane interactions, the balance between hydrophobicity and backbone dipolar interactions in driving ordered peptide aggregation in water and at a hydrophobic-hydrophilic interface, was explored. We found that backbone dipole interactions play a crucial role in driving ordered peptide aggregation, both in water and at hydrophobic-hydrophilic interfaces; while hydrophobicity is more relevant for aggregation in water. A zwitterionic (POPC: 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) and an anionic lipid (POPS: 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine) are used as model lipids for WEPMEM. The addition of head group dipolar interactions in lipids significantly improved structural, dynamic and dielectric properties of the model bilayer. Using WEPMEM and WEPPROM, we studied membrane-induced peptide folding of a cationic antimicrobial peptide with anticancer activity, SVS-1. We found that membrane-induced peptide folding is driven by both (a) cooperativity in peptide self interaction and (b) cooperativity in membrane-peptide interactions. The dipolar interactions between the peptide and the lipid head-groups contribute to stabilizing folded conformations. The role of monovalent ion size and peptide concentration in driving lipid domain formation in anionic/zwitterionic lipid mixtures was also investigated. Our study suggest monovalent ion size to be a crucial determinant of interaction with lipid head groups, and hence domain formation in lipid mixtures. This study reinforces the role of dipole interactions in protein folding, lipid membrane properties, membrane induced peptide folding and lipid domain formation. Therefore, the models developed in this thesis can be used to explore a multitude of biomolecular processes, both at longer time-scales and larger system sizes.

A Visual Analytics Approach to Comparing Cohorts of Event Sequences

Sequences of timestamped events are currently being generated across nearly every domain of data analytics, from e-commerce web logging to electronic health records used by doctors and medical researchers. Every day, this data type is reviewed by humans who apply statistical tests, hoping to learn everything they can about how these processes work, why they break, and how they can be improved upon. To further uncover how these processes work the way they do, researchers often compare two groups, or cohorts, of event sequences to find the differences and similarities between outcomes and processes. With temporal event sequence data, this task is complex because of the variety of ways single events and sequences of events can differ between the two cohorts of records: the structure of the event sequences (e.g., event order, co-occurring events, or frequencies of events), the attributes about the events and records (e.g., gender of a patient), or metrics about the timestamps themselves (e.g., duration of an event). Running statistical tests to cover all these cases and determining which results are significant becomes cumbersome. Current visual analytics tools for comparing groups of event sequences emphasize a purely statistical or purely visual approach for comparison. Visual analytics tools leverage humans' ability to easily see patterns and anomalies that they were not expecting, but is limited by uncertainty in findings. Statistical tools emphasize finding significant differences in the data, but often requires researchers have a concrete question and doesn't facilitate more general exploration of the data. Combining visual analytics tools with statistical methods leverages the benefits of both approaches for quicker and easier insight discovery. Integrating statistics into a visualization tool presents many challenges on the frontend (e.g., displaying the results of many different metrics concisely) and in the backend (e.g., scalability challenges with running various metrics on multi-dimensional data at once). I begin by exploring the problem of comparing cohorts of event sequences and understanding the questions that analysts commonly ask in this task. From there, I demonstrate that combining automated statistics with an interactive user interface amplifies the benefits of both types of tools, thereby enabling analysts to conduct quicker and easier data exploration, hypothesis generation, and insight discovery. The direct contributions of this dissertation are: (1) a taxonomy of metrics for comparing cohorts of temporal event sequences, (2) a statistical framework for exploratory data analysis with a method I refer to as high-volume hypothesis testing (HVHT), (3) a family of visualizations and guidelines for interaction techniques that are useful for understanding and parsing the results, and (4) a user study, five long-term case studies, and five short-term case studies which demonstrate the utility and impact of these methods in various domains: four in the medical domain, one in web log analysis, two in education, and one each in social networks, sports analytics, and security. My dissertation contributes an understanding of how cohorts of temporal event sequences are commonly compared and the difficulties associated with applying and parsing the results of these metrics. It also contributes a set of visualizations, algorithms, and design guidelines for balancing automated statistics with user-driven analysis to guide users to significant, distinguishing features between cohorts. This work opens avenues for future research in comparing two or more groups of temporal event sequences, opening traditional machine learning and data mining techniques to user interaction, and extending the principles found in this dissertation to data types beyond temporal event sequences.