A Combined Functional Data & Mixture Models Approach for Modeling and Classification of Nanomotions
Loading...
Authors
Kweku, David
Issue Date
2021
Type
Thesis
Language
Keywords
Bacterial virulence , Functional Data Analysis , Nanomotion
Alternative Title
Abstract
Nanoscale motion was shown as a new, nonchemical form of detection of life in Kasas et al. (2015). The applications of this technique span from medicine to detection of life in extreme and possibly extraterrestrial environments. One of the importantproblems in this technological approach is the analysis, modeling and classification of the nanomotion data, as it is obtained from the adapted atomic force microscope instrument. The statistical work in this area was very limited up to now, and focused on differentiating between living and inactivated (dead) cells only. In this work we present new results on the statistical analysis and classification of the nanomotion of Bordetella Pertussis. We employed functional data analysis and multivariate statistical analysis methods to develop a classifier for the nanomotion observations as coming from virulent, avirulent or dead bacteria.