A number of computational methods have suggested means by which gene transcription – the process through which RNA is created from DNA – is activated, but there are factors at work that no model has been able to fully explain. In eukaryotes, gene regulation is quite complex, so models have primarily focused on a relatively simple species, Saccharomyces cerevisiae (budding yeast). Because of the inherent complexity in higher species, and even in yeast, a method of identifying transcription factor (TF) binding motifs (specific, short DNA sequences) must be efficient and thorough in its analysis. This thesis shows that a method using the Fast Orthogonal Search (FOS) algorithm to uncover binding motifs as well as cooperatively binding groups of motifs can explain variations in gene expression profiles, which reflect the level at which DNA is transcribed into RNA for a number of genes. The algorithm is very fast, exploring a motif list and constructing a final model within seconds to a few minutes. It produces model terms that are consistent with known motifs, while also revealing new motifs and interactions, and it causes impressive reductions in variance with relatively few model terms over the cell-cycle. / Thesis (Master, Electrical & Computer Engineering) -- Queen's University, 2008-08-21 10:30:24.293

Modelling of human motion is used in a wide range of applications. An important aspect of accurate representation of human movement is the ability to customize models to account for individual differences. The following work proposes a methodology using Hill-based candidate functions in the Fast Orthogonal Search (FOS) method to predict translational force at the wrist from flexion and extension torque at the elbow. Within this force estimation framework, it is possible to implicitly estimate subject-specific physiological parameters of Hill-based models of upper arm muscles. Surface EMG data from three muscles of the upper arm (biceps brachii, brachioradialis and triceps brachii) were recorded from 10 subjects as they performed isometric contractions at varying elbow joint angles. Estimated muscle activation level and joint kinematic data (joint angle and angular velocity) were utilized as inputs to the FOS model. The resulting wrist force estimations were found to be more accurate for models utilizing Hill-based candidate functions, than models utilizing candidate functions that were not physiologically relevant. Subject-specific estimates of optimal joint angle were determined via frequency analysis of the selected FOS candidate functions. Subject-specific optimal joint angle estimates demonstrated low variability and fell within the range of angles presented in the literature. / Thesis (Master, Electrical & Computer Engineering) -- Queen's University, 2008-10-30 01:32:01.606

The recent decade has witnessed great advances in microarray and genotyping technologies which allow genome-wide single nucleotide polymorphism (SNP) data to be captured on a single chip. As a consequence, genome-wide association studies require the development of algorithms capable of manipulating ultra-large-scale SNP datasets. Towards this goal, this thesis proposes two SNP selection methods; the first using Independent Component Analysis (ICA) and the second based on a modified version of Fast Orthogonal Search. The first proposed technique, based on ICA, is a filtering technique; it reduces the number of SNPs in a dataset, without the need for any class labels. The second proposed technique, orthogonal search based SNP selection, is a multivariate regression approach; it selects the most informative features in SNP data to accurately model the entire dataset. The proposed methods are evaluated by applying them to publicly available gene SNP datasets, and comparing the accuracies of each method in reconstructing the datasets. In addition, the selection results are compared with those of another SNP selection method based on Principal Component Analysis (PCA), which was also applied to the same datasets. The results demonstrate the ability of orthogonal search to capture a higher amount of information than ICA SNP selection approach, all while using a smaller number of SNPs. Furthermore, SNP reconstruction accuracies using the proposed ICA methodology demonstrated the ability to summarize a greater or equivalent amount of information in comparison with the amount of information captured by the PCA-based technique reported in the literature. The execution time of the second developed methodology, mFOS, has paved the way for its application to large-scale genome wide datasets. / Thesis (Master, Computing) -- Queen's University, 2010-12-15 18:03:00.208

Due to their complementary characteristics, Global Positioning System (GPS) is usually integrated with standalone navigation devices like odometers and inertial measurement units (IMU). Recently, intensive research has focused on utilizing Micro-Electro-Mechanical-System (MEMS) grade inertial sensors in the integration because of their low cost. In this study, a reduced inertial sensor system (RISS) is considered. It comprises a MEMS grade single axis gyroscope, the vehicle built-in odometer, and two optional MEMS grade accelerometers. Estimation technique is needed to allow the data fusion of RISS and GPS. With adequate accuracy, Kalman filter (KF) fulfills this requirement if high-end inertial sensors are used. However, due to the inherent error characteristics of MEMS devices, MEMS-based RISS suffers from the non-stationary stochastic sensor errors and nonlinear inertial errors, which cannot be suppressed by KF alone. To solve the problem, Fast Orthogonal Search (FOS), a nonlinear system identification algorithm, is suggested in this research for modeling higher order RISS errors. FOS algorithm has the ability to figure out the system nonlinearity with a tolerance of arbitrary stochastic system noise. Its modeling results can then be used to predict the system dynamics. Motivated by the above merits, a KF/FOS module is proposed. By handling both linear and nonlinear RISS errors, this module targets substantial enhancement of positioning accuracy. To examine the effectiveness of the proposed technique, KF/FOS module is applied on RISS with GPS in a land vehicle for several road test trajectories. Its performance is compared to KF-only method, both assessed with respect to a high-end reference. To evaluate navigation algorithm in real-time vehicle application, a multi-sensor data logger is designed in this research to collect online RISS/GPS data. KF/FOS module is transplanted on an embedded digital signal processor as well. Both the off-line and online results confirm that KF/FOS module outperforms KF-only approach in positioning accuracy. They also demonstrate reliable real-time performance. / Thesis (Ph.D, Electrical & Computer Engineering) -- Queen's University, 2012-01-22 01:26:11.477