Using Image Processing Methods to Improve the Detection of Buried Explosive Threats in GPR Data

Current state of the art techniques for landmine detection in ground penetrating radar (GPR) utilize statistical methods to identify characteristics of a landmine response. This research makes use of 2-D slices of data in which subsurface landmine responses have hyperbolic shapes. Various methods from the field of visual image processing are adapted to the 2-D GPR data, producing superior landmine detection results. This research goes on to develop a physics-based GPR augmentation method motivated by current advances in visual object detection. This GPR specific augmentation is used to mitigate issues caused by insufficient training sets. This work shows that augmentation improves detection performance under training conditions that are normally very difficult. Finally, this work introduces the use of convolutional neural networks as a method to learn feature extraction parameters. These learned convolutional features outperform hand-designed features in GPR detection tasks. This work presents a number of methods, both borrowed from and motivated by the substantial work in visual image processing. The methods developed and presented in this work show an improvement in overall detection performance and introduce a method to improve the robustness of statistical classification.

Characterizing the Role of the Previously Undescribed Protein Caskin2 in Vascular Biology

Maintenance of vascular homeostasis is an active process that is dependent on continuous signaling by the quiescent endothelial cells (ECs) that line mature vessels. Defects in vascular homeostasis contribute to numerous disorders of significant clinical impact including hypertension and atherosclerosis. The signaling pathways that are active in quiescent ECs are distinct from those that regulate angiogenesis but are comparatively poorly understood. Here we demonstrate that the previously uncharacterized scaffolding protein Caskin2 is a novel regulator of EC quiescence and that loss of Caskin2 in mice results in elevated blood pressure at baseline. Caskin2 is highly expressed in ECs from various vascular beds both in vitro and in vivo. When adenovirally expressed in vitro, Caskin2 inhibits EC proliferation and migration but promotes survival during hypoxia and nutrient deprivation. Likewise, loss of Caskin2 in vivo promotes increased vascular branching and permeability in mouse and zebrafish models. Caskin2 knockout mice are born in normal Mendelian ratios and appear grossly normal during early adulthood. However, they have consistently elevated systolic and diastolic blood pressure at baseline and significant context-dependent abnormalities in systemic metabolism (e.g., body weight, fat deposition, and glucose homeostasis). Although the precise molecular mechanisms of these effects remain unclear, we have shown that Caskin2 interacts with several proteins known to have important roles in endothelial biology and cardiovascular disease including the serine/threonine phosphatase PP1, the endothelial receptor Tie1, and eNOS, which is a critical regulator of vascular homeostasis. Ongoing work seeks to further characterize the functions of Caskin2 and its mechanisms of action with a focus on how Caskin2-mediated regulation of endothelial phenotype relates to its systemic effects on cardiovascular and metabolic function.

Relative Contributions of Prenylation and Postprenylation Processing in Cryptococcus neoformans Pathogenesis.

Prenyltransferase enzymes promote the membrane localization of their target proteins by directing the attachment of a hydrophobic lipid group at a conserved C-terminal CAAX motif. Subsequently, the prenylated protein is further modified by postprenylation processing enzymes that cleave the terminal 3 amino acids and carboxymethylate the prenylated cysteine residue. Many prenylated proteins, including Ras1 and Ras-like proteins, require this multistep membrane localization process in order to function properly. In the human fungal pathogen Cryptococcus neoformans, previous studies have demonstrated that two distinct forms of protein prenylation, farnesylation and geranylgeranylation, are both required for cellular adaptation to stress, as well as full virulence in animal infection models. Here, we establish that the C. neoformans RAM1 gene encoding the farnesyltransferase β-subunit, though not strictly essential for growth under permissive in vitro conditions, is absolutely required for cryptococcal pathogenesis. We also identify and characterize postprenylation protease and carboxyl methyltransferase enzymes in C. neoformans. In contrast to the prenyltransferases, deletion of the genes encoding the Rce1 protease and Ste14 carboxyl methyltransferase results in subtle defects in stress response and only partial reductions in virulence. These postprenylation modifications, as well as the prenylation events themselves, do play important roles in mating and hyphal transitions, likely due to their regulation of peptide pheromones and other proteins involved in development. IMPORTANCE Cryptococcus neoformans is an important human fungal pathogen that causes disease and death in immunocompromised individuals. The growth and morphogenesis of this fungus are controlled by conserved Ras-like GTPases, which are also important for its pathogenicity. Many of these proteins require proper subcellular localization for full function, and they are directed to cellular membranes through a posttranslational modification process known as prenylation. These studies investigate the roles of one of the prenylation enzymes, farnesyltransferase, as well as the postprenylation processing enzymes in C. neoformans. We demonstrate that the postprenylation processing steps are dispensable for the localization of certain substrate proteins. However, both protein farnesylation and the subsequent postprenylation processing steps are required for full pathogenesis of this fungus.