Rapid Detection of Hepatitis B Virus DNA Level Based on Interval-Point Data Fusion of Infrared Spectra.

Jiaze Chen,Jinfang Ma,Xueqin Han,Yongxin Zhou,Baiheng Xie,Furong Huang,Li,Yuanpeng Li
DOI: https://doi.org/10.1002/jbio.202200251
IF: 3.3899
2022-01-01
Journal of Biophotonics
Abstract:Hepatitis B is an infectious disease cause by the hepatitis B virus (HBV). In recent years, HBV-DNA level clinically gets more attention for its detailed information than other serological markers. Unfortunately, common clinical method for HBV-DNA level detection is limited for its hours consuming. This study combined infrared spectroscopy with machine learning to investigate the feasibility of near-infrared (NIR) and mid-infrared (MIR) spectra for rapid detection of HBV-DNA level. Based on partial least squares-discriminant analysis (PLS-DA) modeling method, the optimal NIR and MIR models and traditional data fusion models were constructed, respectively. Considering inequal weight between interval and point data in machine learning, interval-point data fusion method was used to compare with other traditional date fusion methods. The results of the study illustrate that interval-point data fusion of NIR and MIR spectra combined with PLS-DA modeling can provide a rapid method for HBV-DNA level detection.
What problem does this paper attempt to address?