Search

Rapid changes in computer vision technologies have enabled automatic perspectives for more disciplines that tend to need heavy intervention from human experts. Computational biology, which is the application field of this thesis, is one of those disciplines where computer technologies (software and hardware systems) are applied on cell biology research, drug discovery, and disease diagnosis. The research conducted in this thesis is primarily concerned with automating the analysis of calcium imaging data obtained by two-dimensional (2D) fluorescence microscopy (FM) over living cells. The thesis also presents a theoretical and empirical analysis of the state-of-the-art object detection techniques used in Region based Convolutional Neural Network (R-CNN) and proposes a new R-CNN scheme tailored for cellular object detection in FM data. The analysed images are either individual greyscale images or image stacks of cardiac myocytes stained with DNA markers or calcium indicators. The cells are special cardiac cells found in lung veins, called Pulmonary Vein sleeve Cells (PVC) and Neo-natal Rat Ventricular Myocytes (NRVM) extracted from heart muscle. PVC stains enable the analysis of Calcium signals effect on heart physiology, whereas NRVM images allow autophagy process measuring through accurate cell counting. In the thesis, we demonstrated that automated hotspot detection can be achieved with 79.75% of precision by a two-level segmentation procedure combining thresholding and statistical filtering of cellular regions. We also showed that photobleaching can be corrected by exponential curve fitting and baseline adjustment from normalised calcium traces with respectively a mean square error (MSE) varying between [0.09013, 6.41796] and an overall accuracy of 78.75% for a real-life dataset. Finally, based on the investigation carried over state-of-the-art object detection techniques developed in the past such as the Edge Boxes, the Selective Search, the Objectness Measure, and the Sliding Window paradigm, we demonstrated that a customised R-CNN framework based on a data-driven proposal box generation outperforms with less sampling rate (61 proposals/image) and more ground truth coverage ratio (GTCR of 99.40%).

The contour method has emerged as a promising technique for residual stress measurement in relatively large, thick and complex engineering components. The method involves making a cut in the sample of interest, measuring the subsequent relaxed deformation profile of the cut surface and using this profile to back-calculate the original out-of-plane residual stress field by finite element modelling. The method is based on the theory of elasticity in that the stress relaxation during test specimen cutting is assumed to be entirely elastic. However, when measuring residual stresses of magnitude approaching the material yield strength, plasticity can occur which introduces errors in the measured stress profile. The main aim of this thesis was to develop methods of mitigating and estimating plasticity induced errors in contour method residual stress measurements. Based on the principles of fracture mechanics, an analogy was made between the stress relaxation process and a cracked plate to investigate the origin of plasticity in the contour method. It was demonstrated that that the cut tip stress intensity factor (KT) and the corresponding plastic zone parameters are the most important parameters for characterising plasticity-induced errors in the contour method. Extensive finite element analyses were carried out to understand and control the errors associated with plasticity with a view of improving the accuracy and reliability of the method. The outcomes of this research provide a valuable insight into how accumulation of plasticity for different restraining conditions affects the performance of the contour method. A novel cutting strategy that aims at mitigating plasticity-induced error by controlling the severity of the cut tip stress concentration (i.e. stress intensity factor) during the cutting process has been developed. Furthermore, procedures (correlations) are developed to estimate the plasticity-induced errors in the results of the contour method. Finally guidelines are proposed and applied to a case study for mitigating the errors associated with plasticity in a contour method residual stress measurement.

Data from Meteosat thermal infrared (TIR) imagery are related to rainfall measured directly by gauges at meteorological stations in the Republic of Niger (West Africa), spanning between 0 and 14 °E and 12 and 18 ON. Quantitative physical parameters of 116 clouds from hourly TIR are then examined with the aim of establishing meaningful relations between them and ground observed rainfall.

Calcium oxychloride (CAOXY) formation is a serious deterioration mechanism known to cause joint damage in concrete pavements. CAOXY is the product of a chemical reaction between calcium chloride (CaCl2) or magnesium chloride (MgCl) deicing salts and calcium hydroxide in the cementitious matrix. Currently, the accepted CAOXY threshold at which deterioration is mitigated in cementitious paste is 15 g/100 g paste (15 oz/100 oz paste); however, this limit was developed using flexural strength testing. For this investigation, the current threshold was evaluated using compressive strength and mass change over time in paste specimens exposed to CaCl2. Fly ash is used to replace cement (up to a 50% mass replacement) in order to limit CAOXY deterioration. While compressive strength reduction was observed in all specimens despite fly ash replacement level, the results generally validate the current CAOXY threshold level in cementitious paste determined from flexural strength testing.

The transportation industry is undergoing an unprecedented revolution as researchers in the field expect the adoption of autonomous vehicles (AV) in a not-too-distant future. Even though there is no fully automated vehicle on the road currently, several features of driver’s assistance (e.g., lane departure warning, rear cameras, blind-spot warning) are integrated into most of the recent vehicles. It is therefore fundamental for industry leaders and policymakers to comprehend the state-of-the-art of AV innovation. The main purpose of this study is to assess the current status of AV innovations in the U.S. market. My analysis, based on more than 2,000 patents retrieved from the United States Patent and Trademark Office’s (USPTO) PatentsView database, has five main findings. First, there is a significant increase in autonomous vehicle patents approved by USPTO since 2010. Between 2010 to 2018, the number of patents increased by about 18 folds from 27 to 516. Secondly, in terms of AV innovators, the new entrant high-tech companies are taking over the incumbent automakers in the AV technologies. Third, industries involved in AV innovation have unequal levels of development in different technology sectors and fields. High-tech companies are leading in smart environment technologies. The incumbent automakers had an established predominance in the vehicle platform technologies. Fourth, of all the patents approved by the USPTO, about two-thirds are held by US companies, and one third held by foreign companies primarily from Asia and Europe. Fifth, in the US, California is the epicenter of AV innovation with nearly 40 percent of US patents. Michigan holds 18 percent of the total, given the presence of traditional automobile manufacturers including Ford and GM.

This dissertation evaluates staple food demand in Burkina Faso, specialty coffee sensory attributes and their relation with coffee quality and the effect of material and reputation attributes on specialty coffee quality scores and purchase prices. Chapter One focuses on determining economic and non-economic factors affecting demand for staple cereal commodities in Burkina Faso using the Almost Ideal Demand System (AIDS) model. Results show that millet, maize, sorghum and rice are the main staple cereals. However, we found that millet, maize, and sorghum are necessities while rice is a superior cereal. Disaggregated analyses reveal that households’ consumption patterns change depending on their location, income and education level. The analysis of the evolution of cereal prices shows an overall increase over the period 1996-2009, leaving many people in food insecurity and the country in political instability. Therefore, the government needs to adopt policies that help lower food prices to avoid future food riots. In Chapter Two, we relate specialty coffee cuppers (tasters) and buyers preferences to sensory attributes using descriptive analyses. Principle component analysis (PCA) shows that specialty coffees can be clustered into five groups based on their flavor profiles. The analysis of variance (ANOVA) reveals a significant difference among the means of the five groups. The Tukey-Kramer (TK) method, used for pairwise comparison of the means for the five groups, shows that coffees with sweet, berry and sour flavor profiles have higher mean quality scores and purchase prices, while coffees with brown sugar, other fruit and brown spice flavor profiles have lower mean quality scores and purchase prices. Also, the results reveal that complexity and uniqueness are rewarded with high quality scores and high purchase prices. Since Chapter Two show specialty coffee cuppers and buyers preferences for sensory attributes, the next task is to evaluate the value cuppers and buyers put on those attributes. Therefore, Chapter Three uses the hedonic price technique to measure the value that specialty coffee cuppers and buyers place on material, physical and reputation attributes. The estimation of the quality score equation suggests that material and physical attributes are important determinants of quality. However, reputation attributes have greater explanatory power on quality scores. This result is surprising because the coffee quality score is determined in a blind evaluation, which is independent of reputation attributes. Results of the estimation of the hedonic price show that specialty coffee price is essentially determine by reputation attributes and market conditions such as the number of coffees in the auction. Overall, the analyses show that demand for food depend on economic and demographic variables, sensory attributes affect food quality and price, and reputation attributes explain more of food quality score and price differences than material or physical attributes.

Bacterial fruit blotch (BFB) of cucurbits is caused by Acidovorax citrulli, a Gram-negative seedborne bacterium that can cause up to 100% fruit yield losses in the field. Currently, BFB is a major problem for the cucurbits industry worldwide. Thus far, attempts to identify resistance in cucurbit germplasm for controlling BFB have been unsuccessful. Despite the importance of the disease, little is known about the molecular mechanisms of A. citrulli pathogenicity, due to a lack of molecular tools for studying the A. citrulli/cucurbit interaction. The genomic sequence of A. citrulli strain AAC00-1 has been determined, and the components of type III secretion system have been identified. The goal of this research was to develop molecular tools for studying the BFB disease. Nineteen putative type III effector genes were cloned from two representative A. citrulli strains (AAC00-1 and M6). The distribution of 19 type III effectors among A. citrulli strains, collected worldwide, was studied. A novel Gateway-compatible binary vector was developed for transient expression of A. citrulli type III effectors genes in planta. A set of modified vectors for marker-exchange mutagenesis in A. citrulli were constructed. The model plant species Nicotiana benthamiana was found to be susceptible to A. citrulli, while Nicotiana tabacum was resistance to A. citrulli, so therefore could carry nonhost resistance genes. Two T3S effectors, Aave1548 and Aave2166, triggered water soaking-like cell death in N. benthamiana, but HR-like cell death in N. tabacum. Bacterial mutagenesis and in planta disease assay confirmed that both Aave1548 and Aave2166 have significant virulence contributions to A. citrulli in N. benthamiana plant and melon seeds. Aave2166 encodes a putative acetyltransferase that belongs to the YopJ super family, which is conserved in both animal and plant pathogenic bacteria. Wild type but not the putative catalytic mutant (C232A) of Aave2166 can trigger cell death phenotype in N. benthamiana and N. tabacum. N. benthamiana yeast two-hybrid cDNA library screening using Aave2166 identified six N. benthamiana proteins/peptides which specifically interacted with Aave2166. Further characterization of these Aave2166 interactors may allow us to understand the virulence mechanism provided by Aave2166. The identification of nonhost resistance genes that can recognize Aave2166 and other type III effectors may help to develop novel strategies to control BFB disease of cucurbit.

The lack of efficacy of existing chemotherapeutic treatments of solid tumors is partially attributed to the limited diffusion distance of therapeutics and the low selectivity of anti-cancer drugs with respect to cancerous tissue, which also leads to high levels of systemic toxicity in patients. Thus, chemotherapy can be enhanced through improving anti-cancer drug carrier selectivity and transport properties. Several strains of gram positive (e.g. Clostridium and Bifidobacterium) and gram-negative (e.g. Salmonella Typhimurium and Escherichia coli) bacteria have been shown to possess the innate ability to preferentially colonize tumor tissues. The overall goal of this dissertation is to characterize the transport and sensing of Bacteria-Enabled Drug Delivery Systems (BEADS) in select relevant environments and to investigate the associated underlying principles. BEADS consist of an engineered abiotic load (i.e. drug-laden micro or nano-particles) and a living component (i.e. bacteria) for sensing and actuation purposes. Findings of this dissertation work are culminated in experimental demonstration of deeper penetration of the NanoBEADS within tumor tissue when compared to passively diffusing chemotherapeutic nanoparticles. Lastly, the transport mechanisms that Salmonella Typhimurium VNP20009 utilize to preferentially colonize solid tumors are also examined with the ultimate goal of engineering intelligent and more efficacious drug delivery vehicles for cancer therapy.

Scope and Method of Study: Digitaria exilis (fonio), a wild cereal consumed in West Africa, is rich in chromium, B-glucans and sulfur amino acids, factors known to modulate inflammation and diabetes. However the mechanisms are unknown. We hypothesized that fonio might: a. modulate diabetes pathogenesis through improvement of insulin resistance, glucose metabolism, and reduced inflammation; b. trigger the inflammatory response during infection; c. induce differentiation of immune cells to modulate overall immunity. To test our hypothesis, 3 studies were conducted: In sudy 1, 4-week old KK/HIJ transgenic male mice were fed either the AIN 93 diet or the same diet fortified with 30% fonio or rice flours (n=20). Ten mice per dietary treatment group were sacrificed after 3 or 4 months of feeding. Glucose and insulin tolerance tests (GTT & ITT) were studied by standard techniques. In study 2, non diabetic healthy mice CD1 mice were challenged with bacterial lipopolisaccharides (LPS) and treated with fonio extracts and were sacrified two hours later. To assess the effects of fonio extracts on in vitro secretion by immune cells, spleen cell suspensions (2 x 106/ml) were activated with 2.5 ug/ml LPS for 48 h (study 1 and 2). In study 3, we investigated the mechanisms by which fonio extracts modulate cytokines secretion. For that purpose, monocytic THP1 cells were incubated without and with different concentrations of fonio extracts for 24-72 h. Cell proliferation, viability and differentiation were assessed by trypan blue exclusion test, nitroblue tetrazolium salt reduction, and microscopy, respectively. Cytokines in plasma and spleen cell supernants from the 3 studies were measured by enzyme immune-absorbent.

In this report, we discuss market relations in the cattle and beef sector of the United States by setting up a sequence of optimization decisions taken by cattle feeders (producers, sellers) and meat packers (processors, buyers) to solve for the equilibrium supply and demand proportions in the contract market in a first stage given the respective degrees of risk aversion for the representative producer and processor. Subsequently we derive the impacts of contracts, namely impacts on the spot market price and the processors’ ability to exercise oligopsony power, in a second stage. Using a model for a fixed-price contract, the equilibrium proportions allocated to the contract market by producers and processors correspond to those in Buccola’s model of decisions under risk (1981). This model is extended to solve for the equilibrium in the case of unequal bargaining power between the producer and processor. The residual supply function is used to derive the spot market price through a conjectural variation model and we refer to the prominent Lerner Index for the downstream processor’s oligopsony power. We find that the use of fixed-priced contracts for risk averse producers and processors can serve to reduce the spot market price offered for cattle given the assumption that the residual supply becomes more inelastic. Hence, the processor enjoys higher oligopsony power in the spot market. We run simulations in a spreadsheet model to derive further insights about key parameters, including the risk aversion coefficients, the spot market price variance, the coefficient of expectations, and the bargaining power weights in the contract market. We observe that when the spot market price is expected to be higher than the equilibrium fixed contract price, increases in the producer’s coefficient of risk aversion cause their optimal share of cattle marketed through contracts to increase and hence it contributes to a higher spot market price. On the other hand, when the expected market price is lower than the equilibrium contracted price, increases in the producer’s coefficient of risk aversion decreases the producers’ optimal share of cattle marketed through contracts relative to the processors’ and hence contributes to lower spot market prices. These two results are due to the fact that increasing risk aversion decreases the sensitivity with which optimal portfolios react to changes in the price parameters. By simulating increasing (decreasing) spot market price variance, we observe the equilibrium contracted proportions to increase (decrease) accordingly. In turn, this causes a lower (higher) spot market price. The simulations on the degree of the processors’ bargaining power in the contract market show that increases in the processors’ bargaining power induces higher demand for contracts by processors and lower supply of contracts by producers. This results in a decline in the contract price and an increase in the spot market price. Finally, a brief exercise in a dynamic decision making reveals that a producer will always choose to contract out a proportion of their cattle as long as they are less than 50% certain that, on average, future market prices will be slightly lower than the offered contract price.

Background: Recent studies have found that as many as one in three HIV infected persons may continue to engage in unprotected sex, sexual contact often occurring with seronegative or unknown status partners. While the issue of prevention has received greater attention in developed countries, little is known about the extent to which persons living with HIV/AIDS (PLWHA) might continue to engage in high-risk sexual behaviors in sub-Saharan Africa. Objectives: Propose and test a conceptual model explaining sexual-risk behaviors among PLWHA in terms of cognitive and situational factors. Methods: Cross-sectional study conducted in Abidjan, Cote d’Ivoire (West Africa). The study sample consisted of HIV-positive men and women at various stages of their illness and currently attending one of the largest outpatient clinics for HIV patients. All the data were collected through face-to-face structured interviews. Results: Three hundred forty-nine (349) valid questionnaires were completed. Fifty percent (50%) of the study sample reported one regular partner; 10% reported casual partners and 7% reported multiple partners. Over one-third of the respondents had engaged in high-risk sexual behaviors during the previous 6 months. As posited, two cognitive factors (an internal and external) emerged from the factor analysis process. Although only a small to moderate variance was explained by the model, most of the hypothesized effects were statistically significant: the internal cognitive and the situational factors were inversely related to sexual risk behaviors; the internal cognitive factor also partially mediated the relationship between situational factor and sexual risk-taking. The external cognitive had no significant effect on the outcome. Different models were found for the men and women. Conclusions: The present study indicates that the behavior of a sizable number of HIV patients may continue to increase the burden of the disease at the population. Because access to care is still very limited in sub-Saharan Africa, secondary prevention should be an important part of the routine care for persons living with HIV/AIDS. Moreover, findings suggest that merely encouraging individuals to engage in safe sex behaviors in order to protect others may not be as efficient as appealing to their moral norms and self-confidence for condom use.

Larval black flies were collected monthly from four selected streams in the Great Smoky Mountains National Park between June 1991 and July 1992. One large size (13 m wide Little River), one small (1m wide Tipton Place) and two medium sized streams, one in the sun (6.45 m wide Mill Creek site #1) and one in the shade (6m wide Mill creek site #2) streams and additional creeks in Knox, Monroe, Morgan, Blount counties in eastern Tennessee and Haywood County in western North Carolina will be discussed. Larvae of Simulium tuberosum (Lundstrom) complex, S. venustum/ verecundum Say, S. quebecense Twinn, S.pamassum Malloch, S. impar Davies, Peterson, & Wood, Prosimulium mixtum Syme & Davies, and Stegoptema mutata Malloch were collected from leaves, sticks and rocks submerged in water from different collection sites. S.tuberosum larvae or pupae were present almost year around, while S. venustum/verecundum were present only in October, November, March, April and May in the four streams and rivers. P. mixtum larvae were collected when the water temperature was lowest in December, January, February,March and a few in April. S. quebecense larvae were rare and collected only in March, April,and May. S. pamassum and S. impar were collected from medium size and small size creeks. Stegoptema mutata was found only from Tipton Place and from an unnamed creek in Morgan County. P. magnum was collected mainly in Knox County. Some rare species such as P. arvum, P. rhyzophorum were collected in Knox County and S. loerchae in Sevier County from an unnamed small creek near Pigeon River. Chromosome identification was used to separate sibling species of black flies. Our study of chromosome identification indicated that S. tuberosum species complex consisted of FG/CDE, FG, CKL and A, the S. venustum cytospecies was CC, and S. verecundum cytospecies was AA, Stegopterna mutata would found to be diploid cytospecies. Future research will further address species differentiation with chromosome identification.