Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BHNRC) » Beltsville Human Nutrition Research Center » Food Composition and Methods Development Laboratory » Research » Publications at this Location » Publication #339239

Research Project: Metabolite Profiling and Chemical Fingerprinting Methods for Characterization of Foods, Botanical Supplements, and Biological Materials

Location: Food Composition and Methods Development Laboratory

Title: Develop a Comprehensive Flavonoid Analysis Computational Tool for UHPLC-DAD- HRAM-MSn Data

Author
item ZHANG, MENGLIANG - INTERNATIONAL LIFE SCIENCES INSTITUTE (ILSI)
item Sun, Jianghao
item Chen, Pei

Submitted to: Analytical Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/25/2017
Publication Date: 6/25/2017
Citation: Zhang, M., Sun, J., Chen, P. 2017. Develop a Comprehensive Flavonoid Analysis Computational Tool for UHPLC-DAD- HRAM-MSn Data. Analytical Chemist. 89 (4):7388-7397.

Interpretive Summary: Liquid Chromatography and mass spectrometry methods, especially ultra-high performance liquid chromatography coupled with diode array detection and high resolution accurate-mass multi-stage mass spectrometry (UHPLC-DAD-HRAM/MSn), have become the tool-of-the-trade for profiling flavonoids in foods. However, manually processing acquired UHPLC-DAD-HRAM/MSn data for flavonoid analysis is very challenging and highly expertise. A computational expert data analysis program, FlavonQ-2.0v, has been developed to facilitate this process. The program uses an in-house UV-Vis spectral library and an in-house MS database that contains 5686 flavonoids for the analysis of flavonoids. The program was validated by analyzing data from a variety of samples, including mixed flavonoid standards, blueberry, mizuna, purple mustard, red cabbage, and red mustard green. Accuracies of identification for all samples were above 88%. FlavonQ-2.0v greatly facilitates the identification and quantitation of flavonoids from UHPLC-HRAM-MSn data. It saves time and resources and allows less experienced people to analyze the data.

Technical Abstract: Liquid Chromatography and mass spectrometry methods, especially ultra-high performance liquid chromatography coupled with diode array detection and high resolution accurate-mass multi-stage mass spectrometry (UHPLC-DAD-HRAM/MSn), have become the tool-of-the-trade for profiling flavonoids in foods. However, manually processing acquired UHPLC-DAD-HRAM/MSn data for flavonoid analysis is very challenging and highly expertise-dependent due to the complexities of the chemical structures of the flavonoids and the food matrices. A computational expert data analysis program, FlavonQ-2.0v, has been developed to facilitate this process. The program uses UV-Vis spectra for an initial step-wise division of flavonoids into classes and then identifies individual flavonoids in each class based on their mass spectra. Step-wise identification of flavonoid classes is based on a UV-Vis spectral library compiled from 146 flavonoid reference standards and a novel chemometric model that uses step-wise strategy and projected distance resolution (PDR) method. Further identification of the flavonoids in each class is based on an in-house database that contains 5686 flavonoids analyzed in-house or previously reported in the literature. Quantitation is based on the UV-Vis spectra. The step-wise classification strategy to identify classes significantly improved the performance of the program and resulted in more accurate and reliable classification results. The program was validated by analyzing data from a variety of samples, including mixed flavonoid standards, blueberry, mizuna, purple mustard, red cabbage, and red mustard green. Accuracies of identification for all samples were above 88%. FlavonQ-2.0v greatly facilitates the identification and quantitation of flavonoids from UHPLC-HRAM-MSn data. It saves time and resources and allows less experienced people to analyze the data.