Characterization of cervid skin tissues with chronic wasting disease by Raman spectroscopy and machine learning

Thumbnail Image
Date
2020-12
Authors
Zhu, Binbin
Major Professor
Advisor
Yu, Chenxu
Qiu, Yumou
Rosentrater, Kurt
Greenlee, M. Heather
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Authors
Research Projects
Organizational Units
Journal Issue
Is Version Of
Versions
Series
Department
Agricultural and Biosystems Engineering
Abstract
Chronic wasting disease (CWD) is a contagious neurological disease in cervids that belongs to transmissible spongiform encephalopathies (TSEs). Its spread has threatened the healthy growth of wild and farm-raised deer and resulted in adverse population-level impacts. It also raised concerns over the possibility of infecting human beings like bovine spongiform encephalopathy (BSE). CWD is a prion disease that may take as long as two years for visible signs of the disease to appear. Currently, diagnostic tests approved for official CWD are postmortem tests (immunohistochemistry (IHC) and ELISA) which are not suitable for in vivo diagnosis. Raman spectroscopy offers a potential approach to detect and diagnose CWD rapidly in real time as a first screen onsite. With the Raman spectral data, machine learning algorithms could be utilized to extract meaningful information to differentiate the spectroscopic features that underline the signatures associated with the diseases effectively, even with a low signal-to-noise ratio (SNR) Raman spectral data acquired with a portable Raman spectrometer. In this study, in order to evaluate the effectiveness of Raman spectroscopy on CWD diagnosis, Raman spectra were collected by a Raman microscope as well as a portable Raman spectrometer from cervid skin tissue samples collected from both healthy (i.e., control, CWD-negative) and diseased (i.e., CWD-positive) cervids. The spectral data were classified by two machine learning algorithms, support vector machine and artificial neural network. The results suggested that Raman spectroscopy in conjunction with Machine learning can indeed offer a rapid first screening for CWD, with the highest accuracy of 94.4%. It has the potential to become a useful tool for in-field diagnosis and detection of CWD.
Comments
Description
Keywords
Citation
DOI
Source
Copyright