Browsing by Subject "Algorithms"
Now showing 1 - 5 of 5
- Results Per Page
- Sort Options
Item An Automated Tool for Measuring Aortic Pulse Wave Velocity(2013-01-22) Goel, Akshay; Peshock, Ronald M.; McColl, Roderick; King, Kevin; Whittemore, AnthonyPURPOSE: Aortic Pulse wave Velocity (APV) has been shown to be associated with end organ damage independent of age, sex, and hypertension duration. The purpose of this study is to evaluate an automated approach for computing transit time (Δt) for the measurement of APV as a tool for future investigations and clinical application. METHODS AND MATERIALS: Phase contrast cardiac gated MRI of the aorta in the transverse plane at the level of the pulmonary artery was utilized from the Dallas Heart Study-2 (DHS2), a multiethnic, population-based study of cardiovascular health. A three-step algorithm was used to analyze all 1884 phase contrast MRI studies from the DHS2 central database. The algorithm functions in three key steps: 1) Isolating contours for the ascending aorta and descending aorta using a computer vision technique known as the Hough Transform. 2) Using isolated contours and phase contrast MRI to generate flow curves for the ascending and descending aorta. 3) Computing Δt defined as the time shift between the flow curves in the ascending aorta (AA) and descending aorta (DA), calculated using the half maximum of AA and DA. Fifty of these studies uniformly distributed across all Δt were then randomly selected and manually analyzed with the standard approach utilizing QFlow (v. 4.1.6, Medis) and the corresponding manually derived flow curves were used to compute Δt. The results from the manual analysis using QFlow were compared to results from the automated algorithm using linear regression Bland-Altman difference analysis. RESULTS: The mean Δt in the 1884 studies analyzed with our automated tool was 19.8+/-6.5 ms. In the validation set of 50 studies, linear regression analysis showed an excellent correlation between the automated (A) and manual (M) methods (r=0.97, A = 1.01M-0.885 ms). Bland-Altman difference analysis showed strong agreement with no significant bias (mean difference (A-M) = -0.386 ± 0.768 ms). CONCLUSION: Our automated approach for computing transit time (Δt) for the measurement of APV demonstrates excellent agreement with the standard manual method. These findings suggest this approach could serve as a useful tool for future investigations and clinical application.Item Bringing machine learning to the bedside: focusing clinical decision support with predictive modeling(2019-01-18) Basit, MujeebItem Comprehensive Analysis of Lung Cancer Prognostic Factors(2019-07-29) Wang, Shidan; Gerber, David E.; Xie, Yang; Xiao, Guanghua; Zhan, Xiaowei; Hoshida, YujinLung cancer is the leading cause of death from cancer. It is remarkably heterogeneous in histopathological features and highly variable in prognosis. Analysis of prognostic factor is anticipated to guide clinicians for treatment selection, enhance patient care, and help understanding biological mechanism of tumor progression. To extend current knowledge about lung cancer prognosis, this dissertation analyzed lung cancer prognostic factors in three levels. First, in tumor level, deep learning aided pathology image analysis was used to extract tumor geometry and microenvironment features, upon which an image-based survival prediction model was built and independently validated for lung adenocarcinoma. Second, in patient level, a nomogram was built with demographic and clinical variables for patients with small cell lung cancer. The nomogram was implemented online for public usage. Third, in population level, how facility type and volume affect survival outcome and surgery selection for early stage non-small cell lung cancer was analyzed.Item A New Approach to Optimize a Protein Energy Function on a Folding Pathway Using Gō-Like Potential and All-Atom, Ab Initio Monte Carlo Simulations(2016-01-19) Safronova, Aleksandra; Goldsmith, Elizabeth J.; Grishin, Nick V.; Otwinowski, Zbyszek; Rice, Luke M.Prediction of a protein structure is important for understanding the function of a protein. The process of protein structure prediction employs the approximation of a protein free energy that guides protein folding to the protein's native state. A function with a good approximation of the protein free energy should allow estimation of the structural distance of the evaluated candidate structure to the protein native state. Currently the energy optimization process relies on the correlation between the energy and the similarity to the native structure. The energy function is presented as a weighted sum of components which are designed by human experts with the use of statistical analysis of solved protein strictures. Values of the weights are derived through the procedure that maximizes the correlation between the energy and the similarity to the native structure measured by a root mean square deviation between coordinates of the protein backbone. Two major components are required for a successful ab initio modelling: (1) an effective energy function that discriminates the native protein structure out of all possible decoy structures; (2) an efficient sampling algorithm that quickly searches for the low-energy states. In this dissertation a new method for energy optimization is proposed. The method relies on a fast sampling algorithm and targets successful protein folding. The weights for energy components are optimized on a found with the Gō potential energy fast folding pathway. The Lennard-Jones potential, the Lazaridis-Karplus solvation potential, hydrogen bonding potential are used in the optimization algorithm. The optimized weights successfully predict all α and α/β proteins. The proposed strategy is conceptually different from the existing methods that optimize the energy on solved protein structures. The developed algorithm is a novel concept that allows the optimization of a more complex functional combination of the energy components that would improve the prediction quality.Item An Optical Flow Based Methodology for Visualizing Dynamic Sucellular [sic] Organiztion [sic] Demonstrated Through Profilin and Rho GTPase Microdomains(December 2021) Jiang, Xuexia; Doubrovinski, Konstantin; Jaqaman, Khuloud; Danuser, Gaudenz; Rajaram, SatwikLive cell imaging has enabled the collection of movies of subcellular protein dynamics at a submicron resolution. Statistical time series analysis can greatly expand our understanding of subcellular interactions in minimally perturbed systems. This was previously achieved for the leading edge of migrating cells in select cases. Importantly no strategy existed to simultaneously analyze every subcellular location. Building on existing optical flow based non-linear image registration we developed an approach to remap a migrating cell to a common cell footprint while preserving the characteristics of our signal of interest at a spatial granularity necessary for understanding micron scale biological interactions. This tool enabled us to discover that Profilin fluctuations are organized in living cells. This organization was found to be dependent on cell polarization and actin binding capability. Expanding on this ability to query all subcellular locations, we developed a feature set and feature projection strategy to map molecular biosensor movies of Rho GTPase signaling into micron scale regions of internally consistent signaling dynamics or "microdomains". Microdomains of GTPases match literature descriptions of signaling organization and in an optogenetic study were found to almost precisely match the perturbation footprint.