Abstract:The nutritional value and exposure risk of wild Boletaceae are related to consumers’ health and safety, and influence market stability and trade exports seriously. In this paper, the contents of 12 kind of element contents in eight wild Boletaceae (429 samples in total) were determined and their corresponding infrared spectra data were collected by inductively coupled plasma atomic emission spectroscopy and Fourier transform infrared spectrometry. The content characteristics of elements and edible health risk were analyzed. Meanwhile, partial least squares discriminant analysis and support vector machine (SVM) discriminant models were established using element contents, infrared spectroscopy, low-level data fusion and mid-level data fusion. The classification results were compared. The results demonstrated that wild Boletaceae were rich in Ca, Mg, Na, Zn and other mineral elements, and moderate intake can supplement the nutritional needs of human body. Meanwhile, attention should also be paid to the risk of heavy metal exposure of some Boletaceae. The SVM discriminant model based on mid-level data fusion has the accuracy of training set and test set of 100%, which could be a promising technique to rapidly and accurately discriminate the species of Boletaceae. It can effectively avoid the occurrence of poisoning incidents caused by accidental ingestion. Analyzing the element contents, assessing the health risk and identifying the species of Boletaceae systematically, can provide a reference for the quality safety assessment and the development and utilization of resources.
