文献类型：期刊文献

英文题名：ATR-FTIR spectroscopy with equidistant combination PLS method applied for rapid determination of glycated hemoglobin

作者：Han, Yun[1];Pan, Tao[2];Zhou, Huihui[1];Yuan, Rui[1]

机构：[1]Guangdong Ocean Univ, Dept Math & Appl Math, Haida Rd 1, Zhanjiang 524088, Peoples R China;[2]Jinan Univ, Dept Optoelect Engn, Huangpu Rd West 601, Guangzhou 510632, Guangdong, Peoples R China

年份：2018

卷号：10

期号：28

起止页码：3455

外文期刊名：ANALYTICAL METHODS

收录：SCI-EXPANDED(收录号：WOS:000439315800004)、、EI(收录号：20183005596716)、Scopus(收录号：2-s2.0-85050262959)、WOS

基金：This work was supported by Guangdong Ocean University Scientific Research Start-up Funding for the Doctoral Program (No. R17057).

语种：英文

外文关键词：Mean square error - Least squares approximations - Supersaturation - Fourier transform infrared spectroscopy

外文摘要：A rapid quantification method of glycated hemoglobin (HbA1c) is proposed based on ATR-FTIR spectroscopy in human hemolysate samples. HbA1c, a major clinical diagnostic indicator of diabetes, is a relative content indicator, which was indirectly determined by determining the hemoglobin (Hb) and absolute HbA1c content (HbHbA1c) using ATR-FTIR spectroscopy. The equidistant combination partial least squares (EC-PLS) method was employed in wavelength selection to improve the modeling performance. Modeling samples were divided into calibration and prediction sets 50 times repeatedly to increase the model stability. For Hb and HbHbA1c, 90 and 198 wavenumbers were screened out by EC-PLS, respectively. The predicted values of HbA1c were calculated based on the predicted Hb and HbHbA1c values. The root-mean-square errors (RMSEP) and correlation coefficients of prediction (R-P) were 5.9 g L-1 and 0.945 for Hb, 1.32 g L-1 and 0.931 for HbHbA1c, and 0.67% and 0.918 for HbA1c, respectively, for the validation samples. The results indicated the feasibility of quantification of HbA1c with ATR-FTIR spectroscopy. This method is simpler and quicker than conventional methods. Thus, this process is a potential diabetes screening tool in large populations.