Geographical traceability of germplasm resources of Paris polyphylla var. yunnanensis based on multi-block information integration platform

Abstract

Paris polyphylla var. yunnanensis (Franchet) Handel-Mazzetti as a perennial herb has a medicinal history of about 1800 years in China. Therein, the seeds are the core to ensuring the quality of Chinese herbal medicines. Due to the effect of environmental, climatic, and human factors, the quality of germplasm resources and medicinal materials in different geographical is mixed. Therefore, building a convenient platform for quickly detecting geographic information of P. polyphylla var. yunnanensis germplasm resources is crucial. In this study, multi-block information integration technology (MIR and NIR) combined with pattern recognition methods (PLS-DA and SVM) was adopted to investigate and analyzed the 423 seed samples from six geographical locations in Yunnan Province. The results demonstrated that there were variances in spectral information among germplasm resources from diverse geographical origins; The PLS-DA model based on the LVs dimension reduction strategy and the SVM model based on the VIP dimension reduction strategy could successfully distinguish geographic information of P. polyphylla var. yunnanensis germplasm resources. Among them, the best test set accuracy of PLS-DA model is 96.03%, and the F1 score reaches 96.01%. This study provides a novel and rapid method for selecting excellent germplasm resources of P. polyphylla var. yunnanensis, and ensures that the quality of medicinal materials is controlled by the source.