Rapid Authentication of Dendrobium Species Using Attenuated Total Reflectance Fourier Transform Infrared (ATR-FTIR) and Fourier Transform Near-Infrared (FT-NIR) Spectroscopies with Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA)

Dendrobium has been used as a traditional Chinese medicine for thousands of years known for its effects in nourishing yin, antipyresis, tonifying the stomach, and promoting fluid production. However, different species exhibit significant variations in efficacy and price. Authentication of Dendrobium species involves the quality, safety, and effectiveness of their medicinal use. A fast and reliable method is urgently needed to authenticate these species. In this study, rapid identification models of Dendrobium were developed based on whole-genome resequencing from 54 species, utilizing attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy, Fourier transform near-infrared (FT-NIR) spectroscopy, and mid-level data fusion of these methods. Overall, both ATR-FTIR with 240 variables from wavenumbers 2146.38-485.97 cm-1 and 3029.62-2836.77 cm-1 and FT-NIR with 263 variables from wavenumbers 6078.53-4061.35 cm-1 and 7436.17-6838.34 cm-1 successfully classified the Dendrobium. The orthogonal partial least squares - discriminant analysis (OPLS-DA) models built with 503 variables of mid-level data fusion showed superior differentiation results. The intergroup modeling results of both the training set and the testing set of mid-level data fusion were satisfactory with values of accuracy, precision, sensitivity, and specificity greater than 97.4% and the RMSEP and RMSECV values lower than 0.581 and 0.380, respectively. This study demonstrated the feasibility of combining ATR-FTIR and FT-NIR for the discrimination of Dendrobium, with potential application for other species or other analytical technique combinations. Furthermore, it provides a fundamental basis for higher-level phylogenetic analysis.