University of Maryland, Baltimore County
Not a member yet
17643 research outputs found
Sort by
Autonomous recovery from multi node failures in Wireless Sensor Networks
No full textWireless Sensor Networks (WSNs) are often deployed to serve mission-critical applications in inhospitable environments such as battlefield and territorial borders. Inter-node communication is essential for WSNs to effectively fulfill their tasks. In hostile setups, the WSN may be subject to damage that severs the network connectivity and disrupts the application. The network must be able to recover from the failure and restore connectivity so that the designated tasks can be carried out. Given the unattended operation of the network, the recovery should be performed autonomously. In this thesis we present AuR, an algorithm for Autonomous Repair of damaged WSN topologies. AuR operates in a distributed manner to restore connectivity in the event of multiple node failures. The design principle of AuR is based on modeling connectivity between neighboring nodes as a modified electrostatic interaction based on Coulomb's law between charges. The recovery process is initiated locally at nodes that have lost neighbors. These nodes spearhead the movement in the direction of loss so as to reconnect with other disjoint nodes. The performance of the algorithm is validated through simulatio
Challenges to Inferring Evolutionary Relationships of Closely-related Species: Multilocus Approaches Resolve the Evolutionary History of New World Orioles (Icterus)
No full textMitochondrial DNA (mtDNA) has demonstrated great utility for phylogeography and phylogenetics, but has several limitations and must be corroborated by independent nuclear evidence. Stochastic lineage sorting and introgressive hybridization frequently cause gene genealogies to be incongruent with the species tree and can render even the most strongly supported gene tree misleading. I utilized multiple nuclear introns and several multilocus approaches to investigate the evolutionary relationships between New World orioles within genus Icterus . First, I tested the utility of nuclear introns for inferring species-level phylogenies (Ch. 1). Concatenation analysis of six sex-linked introns yielded a well-resolved phylogeny that corroborated many nodes supported by mtDNA, especially deeper in the oriole tree. Second, I assessed the performance of four species tree algorithms by inferring a phylogeny of a recent radiation of eleven oriole species (Ch. 2). I sequenced seven independent introns from multiple individuals of each species, and found overall agreement between methods and previous data sets. Concatenation and Bayesian concordance analysis were mostly congruent and agreed on eight nodes. One of two coalescent-based methods used (*BEAST) supported six of these eight nodes. I found strong conflict between mtDNA and nDNA regarding the relationships within the well-studied ""northern oriole"" group. MtDNA supported a sister relationship between I. galbula and I. abeillei, whereas nDNA supported a sister relationship between I. bullockii and I. abeillei. Finally, I tested the utility of an extended IM model to examine the divergence population genetics of the three northern orioles (Ch. 3). Multi-population IMa2.0 analysis revealed extensive introgression between I. galbula and I. bullockii. In contrast, there was no gene flow between I. bullockii and I. abeillei, suggesting that their close nDNA relationship is not due to extensive introgression. However, previously undocumented substantial introgression between I. galbula and I. abeillei suggests that their close mtDNA relationship is likely due to mtDNA introgression and replacement in I. abeillei, causing the mtDNA tree to be misleading. Increased documentation of introgression in nature highlights the need for methods that can infer both trees and population parameters so that we can account for gene flow when inferring the evolutionary history of closely-related species
Privacy Preserving PCA on Distributed Bioinformatics Datasets
No full textIn recent years, new bioinformatics technologies, such as gene expression microarray, genome-wide association study, proteomics, and metabolomics, have been widely used to simultaneously identify a huge number of human genomic/genetic biomarkers, generate a tremendously large amount of data, and dramatically increase the knowledge on human genomic/genetic information, thus significantly improving biomedical research. However, these exciting advances in bioinformatics do come with a drawback: the increasingly richer human genomic/genetic data contains sensitive private information, such as genetic markers, diseases, etc., which may further lead to the discovery of the individual's race, family, or even identity. Therefore, privacy is an important issue when dealing with bioinformatics data. This is further exacerbated when multiple data providers try to collaborate with each other. This dissertation presents a set of novel approaches for Privacy Preserving Principal Component Analysis (PP-PCA) computations on genomic data from several distributed parties. The approaches allow data providers to collaborate together to identify gene profiles, biomarkers, and possible new pathways from a global viewpoint, and at the same time protect sensitive genomic data from possible privacy breaches. Based on our approaches, we further provide a PP-PCA gene clustering framework and workflow that includes two types of roles: data providers and a trusted center. Within this mechanism, scenarios of horizontal, vertical, and mixed partitioning are covered. Under the horizontal partitioning scenario, distributed genomic datasets can be processed for global PCA gene clustering analysis with privacy protection. Furthermore, compared to the results from a centralized scenario, the results calculated from distributed partitions using our mechanism maintain 100% accuracy. Experiments on five genomic datasets are conducted, and the results show that our framework produces exactly the same results as from merged datasets. In the vertical partitioning scenario, two different methodologies are employed: Collective Principal Component Analysis (CPCA) and Repeating Principal Component Analysis (RPCA). CPCA requires local sites to transmit a sample of original data to a Trusted Center Site (TCS). CPCA can be applied to datasets each having a different number of columns. The RPCA approach requires that all local sites have the same or similar number of columns, but releases very little information of original datasets. Experiments on five genomic datasets show that both CPCA and RPCA approaches maintain very good accuracy compared with a centralized scenario. Under the mixed partitioning scenario, the more generic situation, multiple conditions are discussed, and the conditions of ""Vertical Partitioning with Extra Rows"" (VPER) and ""Horizontal Partitioning with Extra Columns"" (HPEC) were identified as the valuable and practical types of mixed partitioning scenario. Both CPCA- and RPCA-related methodologies are applied to the VPER condition, and horizontal partitioning related method is applied to the HPEC condition. Overall, this dissertation offers multiple approaches to build a framework to handle multiple situations on distributed PCA gene clustering, and experimental results show it could obtain accurate global results and preserve data privacy
Protein-Protein Interactions of Rob and SoxS and their role in stress defense response systems in Escherichia coli.
No full textEscherichia coli Rob, SoxS and MarA are monomeric members of the AraC/XylS family of transcription activators, possessing two helix-turn-helix motifs that are used for DNA binding and transcription activation. SoxS and MarA are synthesized de novo in response to superoxide stress and salicylate. Rob is unique, being constitutively expressed at 5,000-10,000 molecules per cell, but maintained in an inactive form by being sequestered into 3-4 intracellular aggregates. Dipyridyl (DIP) and bile salts (DEC) induce Rob activity, by dispersing the aggregates into monomers that activate transcription of 40-50 genes that mediate the cell's defense response against the above stresses. Rob's activity is controlled by sequestration-dispersal (S-D), a new mechanism of induction where under normal conditions Rob resides in an inactive, sequestered state, primed to be dispersed in response to environmental stress. Our goal is to determine how Rob is sequestered and whether a cytoplasmic factor converts dispersed, active Rob into its inactive, sequestered form. In this study, Rob was tagged with Protein A or His₁₀ and protein complex immunoprecipitation was carried out to isolate complexes between Rob and the putative factor (Chapter IV). Potential protein partners of Rob were then identified by mass spectrometry. In addition, we conducted alanine-scanning mutagenesis of Rob's C-terminal domain (CTD), the portion of the protein shown previously to be necessary and sufficient for S-D (Chapter II). In previous work, several regions important in aggregation and induction were identified using quad alanine-scanning mutagenesis. In this work, single alanine mutants of these regions were made in order to identify the amino acids important to the formation or maintenance of the aggregates or to inducer binding. Several mutant phenotypes of Rob were observed: constitutive, uninducible, partially constitutive/inducible and super-aggregated/inducible. Furthermore, direct protein-protein interactions between Rob and RNAP were identified at class II promoters (Chapter V). Moreover, we determined Rob's orientation at class II promoters and demonstrated that Rob occludes σ⁷⁰ region 4 from binding the -35 hexamer at class II promoters (Chapter V). Finally, Rob was shown to form in vivo binary complexes with RNAP, which supports pre-recruitment as Rob's mechanism of transcription activation (Chapter III)
Multi-Scale Analysis of Observations of Tropical Cyclones with Applications to High-Resolution Hurricane Modeling
No full textTropical cyclone numerical models, a critical tool to forecasters, have been run at resolutions of around 9-30 km in operational centers until recently. It is currently possible to run in the range of 1-4 km resolution, which may allow a model to resolve small-scale dynamical processes critical to tropical cyclone intensity. A 3km version of the NCEP HWRF model is developed for that purpose and its competitive track and intensity forecasting abilities are demonstrated. To determine if the small scales are resolved correctly, a statistical framework for comparison to observations of small-scales is developed. The standard definition of a model's forecast intensity is examined, and found to have a systematic, resolution-dependent bias. A database of TRMM overpasses of over eight hundred tropical cyclones is produced and used to show a relationship between storm-scale cloud top temperature and storm wind intensity. However, all storms, regardless of strength, produce near-tropopause cloud tops, and storms undergoing rapid intensification (RI) tend to have higher cloud tops than non-RI storms. In an analysis of in-situ wind data, vertical wind is shown to be scale-invariant, with no correlation beyond, nominally, 2 km scales. This new framework for comparison is used to show that model's cloud tops have the right relationships with intensity and intensification, but that downdrafts are weak and rare. Model ""spin-up"" issues are seen: in the first six hours, some storms rapidly gain fine-scale 3 km resolution wind maxima that hurt the forecast and others weaken uniformly at all resolutions. In addition, a model bug is found in this and operational HWRF: all microphysics type fractions are discarded when the nest moves. Overall, the research presented in this demonstrates the value of statistical diagnostics for high-resolution models. In addition, this research presents a framework for a deeper investigation of tropical cyclone small-scale dynamics
[EB 2011, Business Meeting]
No full text[Title supplied by cataloger]AAA President Kathy Jones (center) with outgoing Board members Philip Brauer, Joseph Sanger, Erica Perryn and David Burr (L-R) at the 2011 Business Meeting