Honors Program
Document Type
Thesis
Date of this Version
8-2024
Citation
Schommer, JB. (2024). The Use of Fourier-Transformed Infrared Spectroscopy with an Infrared Biotyper to Establish Classifiers for Major Serogroups of Moraxella Bovis Based on Spectral Profiles. Undergraduate Honors Thesis. University of Nebraska-Lincoln.
Abstract
A classifier tool was created to identify major serogroups of Moraxella bovis based on spectral profiles collected from bacterial cells through the use of Fourier transformed infrared spectroscopy. M. bovis strains were selected from a group of previously sequenced and genotyped isolates, which provided the pilin sequence type for each isolate. The obtained spectroscopic profiles were analyzed through artificial neural networks to create a classifier tool that could identify the PilA type for unknown strains. Once a sufficient spectral library was obtained, the isolates were split into a training and a validation group, each containing spectra from isolates of five major PilA serogroups. Once trained, the classifier model was tested on the validation group where it displayed 94.6% overall accuracy at identifying isolates with the correct PilA serogroup and a Kappa statistic of 0.505, which reflects moderate agreement.
Included in
Gifted Education Commons, Higher Education Commons, Large or Food Animal and Equine Medicine Commons, Veterinary Infectious Diseases Commons, Veterinary Microbiology and Immunobiology Commons
Comments
Copyright Jordyn Schommer 2024