An Intelligent Extraction Approach of Feature Information from Three-Dimensional DAD Chromatogram for Integrated Quality Control of Traditional Chinese Medicine: Gardenia Jasminoides root as an example

Traditional Chinese medicines (TCMs) fingerprint is a conventional strategy for the holistic quality control of TCMs, which has the characteristics of holism and ambiguity. However, the fingerprint of TCMs at the current stage generally adopts a single wavelength or few wavelength, lacking the effective utilization of diode-array detector (DAD) chromatogram data. This study proposed an intelligent extraction approach of feature information from three-dimensional DAD chromatogram to establish a novel Bar-Form-Diagram (BFD) for integrated quality control of TCMs. BFD was automatically established by the chromatographic and spectral information of a complex hybrid system in a DAD chromatogram. It covered the peak areas of the target compositions at the optimal absorption wavelength. Taking 27 batches of Gardenia Jasminoides root as samples, BFD combined with chemometrics were applied for assessing the quality of samples completely, which improved the accuracy of origin classification with hierarchical cluster analysis, principal component analysis, soft independent modeling of class analogy and orthogonal partial least squares discriminant analysis. Single wavelength fingerprint and BFD used 23 and 38 common peaks as variables respectively, and the adjusted rand index results of the single wavelength and BFD were 0.559 and 0.819, respectively. Compared with ergodic methods of each single-wavelength, the peak recognition method in this study improved the operation speed from 180s to 4s and the computational complexity. The established BFD approach performed more abundant characteristic information of chemical components of TCMs and more accurate classification ability of the origin, and it had great advantages in the overall quality control of TCMs.

You have access to this article

Please wait while we load your content... Something went wrong. Try again?

留言 (0)

沒有登入
gif