Application Of Raman Spectroscopy In Serum Diagnostic Test Of Hepatitis B Virus

Objective:To study the screening of Hepatitis B virus（HBV）infection by serum detection using Raman spectroscopy.Methods:1.Select 20 cases of HBV-infected sera and non-HBV-infected sera,respectively,and use Raman spectrometer to detect them without using wavelengths to select the best experimental conditions.2.Under the above optimal conditions,500 serum samples from the study group and 500 serum samples from the control group were selected for testing.The measured spectrum was firstly adapted to the adaptive iterative re-weighted penalized least square method.squares（airPLS）to subtract the fluorescence background from the Raman spectrum,then use Principal Component Analysis（PCA）to extract the principal component features of the processed Raman spectrum,and then use the Support Vector Machine（SVM）For modeling and prediction,Particle Swarm optimization（PSO）algorithm replaces traditional grid search to optimize SVM parameters.Finally,a diagnostic model of HBV infection was established.3.Double-blind method was used to detect 600 cases of serum to verify the efficacy of the test according to the established model and the application of fluorescent quantitative polymerase chain reaction（PCR）technology.4.PCR-reverse dot hybridization was used to select 151 cases of HBV infected sera to determine whether rtV204 gene mutation.After Raman spectroscopy was used to detect 151 serum samples,the measured spectrum was deducted by airPLS and the partial least square method(Partial Least Squares（PLS）feature extraction,SVM modeling and prediction,to establish a diagnostic model of HBV infection and rtV204 gene mutation.Results:1.The optimal experimental conditions for this experiment are 532nm.2.It is found in the Raman spectrum that there is a difference between normal human Raman peaks and patients with hepatitis B virus at509,957,1002,1153,1512,1648,2155,2305cm^-1.Using airPLS-PCA-PSO-SVM to establish a diagnostic model for HBV-infected serum,the accuracy was 93.1%,the sensitivity was 100%,the specificity was 88%,and the area under the ROC curve was0.942.3.In this study,we found that the Raman peaks of HBV-infected and rtV204 gene mutant serum samples and unmutated serum samples were 877,957,1002,1153,1280,1447cm^-11 There are great differences at 1,510,and1654cm^-1.The airPLS-PLS-PSO-SVM model has a diagnostic accuracy of 84.78%,a sensitivity of 68.75%,a specificity of93.33%,and an area under the ROC curve of 0.847 for HBV infected sera with rtV204gene mutation.4.Double-blind method was used to verify the effectiveness of the two groups of tests,the first group:sensitivity:87%,specificity:92%,KAPPA value:0.79;the second group:sensitivity:80%,specificity:79%,KAPPA value:0.59.Conclusion:This preliminary study shows that serum Raman spectroscopy combined with airPLS-PCA-PSO-SVM model can be used for screening for hepatitis B virus.The airPLS-PLS-PSO-SVM model has very good sensitivity and specificity for the diagnosis of rtV204 gene mutations in HBV.