Mid-infrared spectroscopy coupled with chemometrics for quantitative determination of biomaterial activity

Abstract

In view of the complicated operation and time-consuming of traditional detection methods for biomaterial activity, a method based on the mid-infrared (MIR) spectroscopy coupled with chemometrics methods was proposed to achieve the non-destructive and rapid determination of the biomaterial activity. The spectral preprocessing methods were applied to reduce the noise and scattering effects in the spectral analysis, which can effectively improve the performance of biomaterial activity determination model. The characteristic wavelengths of the preprocessed spectrum were extracted by the successive projections algorithm (SPA). And the determination performance of different regression algorithms for biomaterial activity was compared and discussed. Results demonstrated that the model of SPA combined with partial least squares regression (PLSR), which exhibited better potential for the determination of biomaterial activity with higher coefficient of determination (R2) and lower root mean square error of prediction (RMSEP). The R2 acquired by the optimal model for three biomaterials are 0.9622, 0.9949, and 0.9936, respectively. And the RMSEP were 0.0611, 0.0186, and 0.0193, respectively. Therefore, the SPA-PLSR established by MIR is expected to replace the traditional methods for quantitative determination of biomaterial activity.