Fuente: PubMed "wine" J Pharm Biomed Anal. 2025 Dec 8;270:117312. doi: 10.1016/j.jpba.2025.117312. Online ahead of print.ABSTRACTAconitum Traditional Chinese medicine, represented by Chuanwu (Aconiti Radix), Caowu (Aconiti Kusnezoffii Radix), Fuzi (Aconiti Lateralis Radix Praeparata), and their processed products ZhiChuanwu (Aconiti Radix Cocta), ZhiCaowu (Aconiti Kusnezoffii Radix Cocta), and ZhiFuzi, constitute a class of Chinese herbs characterized by a "dual nature of toxicity and efficacy." They possess pharmacological effects such as dispelling wind, eliminating dampness, and dispersing cold to relieve pain. Due to their high toxicity, the processed forms-ZhiChuanwu, ZhiCaowu, and ZhiFuzi-are predominantly used in clinical practice. ZhiFuzi is a key herb for warming Yang. ZhiChuanwu specializes in treating cold-bi syndrome (painful obstruction due to cold), and ZhiCaowu excels in pain relief. Therefore, the correct classification of these toxic Aconite materials is crucial for ensuring their quality control and appropriate clinical application. This study proposes a data fusion and multivariate analysis strategy based on near-infrared spectroscopy (NIR), electronic nose (E-nose), and high-resolution mass spectrometry (HRMS) for classifying these toxic Aconite herbs. Traditional chemometric modeling based on a single data source was found incapable of correctly classifying the three types of herbs. However, chemometrics combined with data fusion strategies enhanced the performance of the classification models. Furthermore, to identify the optimal combination of analytical modeling methods, the performance of different classification algorithms under various data training strategies was compared. PCA-Linear Discriminant Analysis (PCA-LDA) and PLS-DA, when combined with the feature-level fused dataset (NIR-E-nose-MS), yielded the best classification results. Overall, the classification strategy established in this study holds significant potential for classifying toxic Aconite medicinal materials.PMID:41391330 | DOI:10.1016/j.jpba.2025.117312