Proteome profiling is the identification and quantitation of all proteins in biological samples. An important application of proteome profiling that has received much attention is clinical proteomics, a field that promises the discovery of biomarkers that will be useful for early diagnosis and prognosis of diseases. While clinical proteomic methods vary widely, a common characteristic is the need for (i) extraction of proteins from complex biological fluids and (ii) extensive biochemical processing (reduction, alkylation and enzymatic digestion) prior to analysis. However, the lack of standardized sample handling and processing in proteomics is a major limitation for the field. The conventional macroscale manual sample handling requires multiple containers and transfers, which often leads to sample loss and contamination. For clinical proteomics to be adopted as a gold standard for clinical measures, the issue of irreproducibility needs to be addressed. A potential solution to this problem is to form integrated systems for sample handling and processing, and in this dissertation, I describe my work towards realizing this goal using digital microfluidics (DMF). DMF is a technique characterized by the manipulation of discrete droplets (100 nL – 10 L) on an array of electrodes by the application of electrical fields. It is well-suited for carrying out rapid, sequential, miniaturized automated biochemical assays. This thesis demonstrates how DMF can be a powerful tool capable of automating several protein handling and processing steps used in proteomics.

Diagnostic ultrasound is a non-invasive image modality commonly used to diagnose multiple diseases. Sonographers and physicians have to devote a substantial amount of time learning how the combination of the various parameters of the ultrasound system affects the resulting ultrasound information among its different modalities. The primary objective of this thesis is to create a mechanistic ultrasound simulation method that could achieve near real-time performance to teach and train sonographers and physicians. A major bottleneck for achieving real-time performance with available tools was the transducer field simulation method that uses the impulse response technique. To address this problem a new simulation approach was developed. This project was realized in a three-phase approach, specifically to simulate spectral Doppler. In first place, it was shown that is possible to mechanistically simulate the sample volume power distribution using a novel method based on an array of point sources to represent the transducer geometry. Secondly, by comparison with the Field II results, it was demonstrated that the time-domain signal could be closely reproduced using point sources. Finally, by treating the array of point sources as point receivers, the received signal was compared with the results from Field II, and again, good agreement was achieved. Simulation results were compared against the standard simulation method for a number of examples involving steady and pulsatile flow, for which the spectrograms were compared against Field II. Also presented are preliminary results obtained using the point source approach to simulate B-mode images. As well, methods are described for generating Doppler spectrograms from the results of computation fluid dynamics velocity fields obtained in realistic arterial geometrical models. It is pointed out that the successful simulation of the time domain signal opens the possibility for real-time simulation of other ultrasound modes.

Engineered protein systems have been created to impart new functions, or “re-program” mammalian cells for applications including cancer and HIV/AIDS therapies. The successful development of mammalian cells for re-programming will depend on having well-defined, modular systems. Migration is a particularly important cell function that will determine the efficiency and efficacy of many re-programming applications in vivo, and Rho proteins are responsible for regulation of cell migration natively. While there have been several reports of photo-activated Rho proteins, no strategy has been developed such that Rho proteins and cell migration can be controlled by a variety of extracellular stimuli that may be compatible with signaling in large organisms. Here, several methods are described for engineering Ca2+-sensitive Rho proteins so that the large, natural toolbox of Ca2+-mobilizing proteins can use the Ca2+ intermediate to activate Rho proteins in response to a variety of exogenous stimuli, including chemicals, growth factors, and light. First, an unreported calmodulin binding site was identified in RhoA. This knowledge was used to create a tandem fusion of RhoA and calmodulin that mediated Ca2+-sensitive bleb retraction in response to a variety of Ca2+-elevating chemicals. Ca2+-mobilizing modules including channelrhodopsin-2 and nicotinic acetylcholine receptor α4 were used for light- and acetylcholine-dependent bleb retraction. Second, a more robust morphology switch was created by embedding a calmodulin binding site into RhoA to enable Ca2+-responsive bleb formation. A wider range of Ca2+-mobilizing modules were also used here including LOVS1K/Orai1 and vascular endothelial growth factor 2. Combining Ca2+-mobilizing and Ca2+-responsive modules increased amoeboid-like cell migration in wound closure and transwell assays. Finally, the embedded peptide design was applied to Rac1 and Cdc42 to enable control of new morphologies and migration modes. The modular Ca2+ control over Rho proteins developed here is an important contribution to cell re-programming because it shows that control over cell migration can be rewired in a way that is flexible and tunable.

Protein transduction domains (PTDs) are short peptide sequences that can transport wide varieties of cargo across cell membranes. This study assessed the transduction ability of fusion proteins containing optimised variants of the PTD from HIV-1 transactivator of transcription (Tat). Uptake of Tat-PTDs was determined by fluorescent microscopy using the fluorescent protein Venus as a tag, and also by using fusion proteins containing caspase-7 and RhoA bound to Tat-PTD. Upon entering the cytosol the latter two induce apoptosis and the formation of cytoplasmic extensions, morphological changes easily observed by microscopy. It was found that PTDs with two, three or four sequential Tat-PTD domains could bind to the surface of two of the five cell lines tested. Fluorescent microscopy, however, indicated that the fluorescent constructs remained on the cell surface. As well, PTDs bound to caspase-7 or RhoA did not induce any visible morphological changes in the cells.

The work in this thesis was concerned with the safety evaluation of a fully powered mechanical patient lift for bariatric patients. A working prototype of this system, called RoboNurse, was designed and manufactured at iDAPT technology team at Toronto Rehabilitation Institute. There are currently no lifting technologies similar to RoboNurse in the healthcare industry. The methods that are used to evaluate the system included: 1) Series of mechanical tests to evaluate the static strength and stability of the design 2) Computer simulations to evaluate the dynamic stability of the system and 3) Failure mode and Effects Analysis (FMEA) and Fault Tree Analysis (FTA) as risk analysis tools. These techniques helped to perform thorough and systematic evaluations on the system and its components. This study significantly assisted in understanding the problems associated with the current design prototype and provided the necessary resources and guidelines for the future generations of this technology.

This study aimed to apply human factors methods to identify potentially unsafe aspects of a radiation therapy delivery system at Princess Margaret Hospital, and to provide recommendations accordingly. Analyses were conducted to examine the workflow, work environment and user interfaces involved in the treatment process. Based on findings from these analyses, components of the user interface were redesigned to address some of the issues found. Sixteen radiation therapy students were then used to experimentally evaluate the redesigned interface through a usability test. Compared to the current interface, the error rates of two common errors were significantly lower, and the average task completion time was significantly shorter when the redesigned interface was used. Results from a post-test questionnaire also indicated a high degree of satisfaction with the redesigned interface. Therefore, human factors methods can be applied to evaluate and design radiation therapy systems for improved error rates, efficiency and user satisfaction.

Swallowing dysfunction, or dysphagia, can result from any structural or neurological impairment that affects the swallowing mechanism. Accelerometry is a non-invasive technique that may be used to measure the epidermal vibration signals on a patient’s throat during swallowing. The purpose of this study was to map out the trajectories of the hyoid bone and arytenoids during swallowing to determine if a relationship exists between their trajectories and the signals obtained from an accelerometer placed on a participant’s neck during a scheduled videofluoroscopic swallowing study. Participants included 43 adult patients with neurogenic dysphagia. A software program was created to track the desired anatomy on the videofluoroscopic videos frame-by-frame. The data from the software was then compared to the integrated acceleration data multiple linear regression. The results show that the hyoid and larynx are not always correlated during swallowing and further suggest hyolaryngeal excursion as the primary physiological source of accelerometry.

The vertebral column is the most common site of skeletal metastatic development secondary to breast cancer. Multiple clinical treatments are available for spinal metastasis, including systemic bisphosphonates and radiation therapy, however the success of current treatment approaches varies considerably. Alternative treatment strategies for spinal metastatic destruction must be aimed at both reducing tumor burden and restoring mechanical stability. Photodynamic therapy (PDT) has been shown to be successful at destroying osteolytic lesions in preclinical models of breast cancer spinal metastasis. However, the clinical feasibility of PDT for spinal metastasis is dependent on its potential effects on the structural integrity of vertebral bone. This thesis aims to determine the effects of PDT alone and in combination with bisphosphonate therapy on the structural architecture and mechanical properties of healthy and metastatically involved vertebrae. PDT was shown to have a positive effect on vertebral bone structure, alone and in combination with previous bisphosphonate therapy.

Driving is a cognitively complex task, yet the areas of the brain involved in driving are not well understood. This thesis investigates the neural correlates of driving using functional magnetic resonance imaging and driving simulations with custom built driving hardware (steering wheel and foot pedals). The feasibility of driving in fMRI is first investigated, and a methodology is described to minimize head motions. Next, the functional neuroanatomical correlates of driving tasks of varying complexity are explored. Simple tasks such as straight driving activate areas of the brain related to motion, spatial navigation and coordination. Complex tasks are found to recruit additional areas of the brain, including areas of higher cognitive function such as the prefrontal cortex. Developing a better understanding of the areas involved in different driving tasks is an important first step in understanding the neural basis of driving skill and safe driving.

Administration of medication directly to diseased lungs reduces adverse systemic side effects. For cystic fibrosis, jet nebulizers are the standard aerosol delivery system since they can aerosolize drugs that require relatively large volumes of liquid. Selection of the appropriate nebulizer for a given drug is crucial to ensure delivery of the therapeutic dose. This selection, ideally, requires knowledge of the pulmonary drug deposition (PDD). The gold standard for accurately measuring PDD is nuclear medicine techniques, which exposes the subject to radiation and therefore cannot be used repeatedly to test multiple devices. An alternative is to characterize the nebulizer using in vitro experiments and estimate the device’s in vivo performance. However these techniques are time-consuming and can only collect data for one breathing pattern and drug-device combination. Therefore this study is to formulate mathematical models for jet nebulizers that can estimate PDD based on the drug-device combination and patient’s breathing patterns.