Predicting the contents of polysaccharides and its monosugars in Dendrobium huoshanense by partial least squares regression model using attenuated total reflectance Fourier transform infrared spectroscopy

An attenuated total reflectance Fourier transform infrared spectroscopy method combined with partial least square regression was established to rapidly quantify the total polysaccharide and its major monosugars in Dendrobium huoshanense. The results showed that the optimal predictive methods were the models established by the infrared spectroscopy files pretreated by standard normal transformation combined with the second derivative; the accuracy of the models selecting wavelength regions based on the assignments of Fourier transform infrared spectroscopy signals of polysaccharide was obviously higher than the models selecting wavelength regions suggested by Thermofisher Quantity Analyst software and other possible wavelength region selecting modes. The external validation and the complete external validation confirmed the robustness and reliability of the developed attenuated total reflectance Fourier transform infrared model. Our study might provide an efficient technique tool for the rapid, green, low-cost, and nondestructive quantification of the total polysaccharide and the main monosaccharides in Dendrobium huoshanense and other rich-in-polysaccharide plant food or medicines.