Tumor genesis is accompanied by glycosylation of related proteins. Glycoprotein is usually regarded as a tumor marker since glycoproteins are consumed remarkably more by the cancer cells than the normal ones. In this paper, the terahertz time-domain attenuated total reflection (ATR) technique is applied to inspect the glycoprotein solution from a concentration gradient of 0.2 mg/ml to 50 mg/ml. A significant nonlinear relationship between the absorption coefficient and the concentrations has been discovered. The influence of the dynamical hydration shell around glycoprotein molecules on the absorption coefficient is discussed and the phenomenon is explained by the concepts of THz excess and THz defect. In order to identify glycoproteins, features are obtained by composite multiscale entropy (CMSE) method and clustered by the K-means algorithm. The results indicate that features extracted by the CMSE method are better than the Principal Component Analysis (PCA) method in both specificity and sensitivity of recognition. Meanwhile, the absorption coefficient and dielectric loss angle tangent are more suitable for qualitative identification. Research shows that the CMSE method has important directive significance for analyzing glycoprotein terahertz spectroscopy. And it has the potential for glycoprotein related tumor markers identification using terahertz technology in medical applications.
Keywords: Absorption coefficient; Composite multiscale entropy (CMSE); Dielectric loss tangent; Glycoprotein; Terahertz time-domain attenuated total reflection spectroscopy; Tumor marker identification.
Copyright © 2018 Elsevier B.V. All rights reserved.