Linear discriminant analysis for detection of salivary NS1 from SERS spectra

Abstract

Surface Enhanced Raman spectroscopy (SERS) is an enhanced technique of Raman spectroscopy. It produces Raman spectra with higher intensity peak than the conventional Raman spectroscopy. Using SERS, detection of low concentration of NS1 in saliva seems promising. NS1 protein is an early biomarker for viral diseases caused by flavivirus such as Dengue fever, Yellow fever, Japanese Encephalitis etc. NS1 is Raman active, hence produces unique Raman spectrum with NS1 characteristic peaks. However, at low concentration, the intensity of NS1 characteristic peak is too subtle to be detected visually. Hence, signal processing techniques are crucial to extracting the characteristic peak from the spectra. Linear Discriminant Analysis (LDA) is a signal processing technique which can be used as a classifier. Here, LDA is coupled with Principal Component Analysis (PCA) to classify Raman spectra of saliva with and without NS1. Using k-fold cross validation technique, the highest accuracy achieved is 98.4% with the corresponding sensitivity of 96.9%, precision of 100% and specificity of 100%. The highest performance is achieved with 70 retained principal components as proposed by Cumulative Percentage Variance (CPV) stopping criterion.