MS-FINDER and SIRIUS for phenolic compound identification from high-resolution mass spectrometry data

Abstract

LC-HR-MS/MS is the predominant analytical technique in phenolic compound (PC) research. However, the manual interpretation of mass spectra is a heavy nontrivial time-consuming task and depends on mass spectrometry and phenolic compounds fragmentation deep knowledge. We think this manual approach should be partially translated into a practical software that allows users to perform such complicated analyses. In silico fragmentation software have been tested for small molecule identification, MS-FINDER and SIRIUS stood out at identification contests and challenges. We evaluated both software to identify PC from two data categories: 1^st MS/MS spectra from 18 phenolic compound standards (PCS) and 2^nd phenolic compounds from 8 food samples (FPC) (coffee, green tea, cranberry juice, grape juice, orange juice, apple juice, soy extract and parsley extract). MS-FINDER and SIRIUS were able to correctly identifymore than 90% of the PCS by LC-HR-MS/MS. The main FPC were also correctly identified by MS-FINDER (70%) and SIRIUS (38%). We highlight that these software were unable to differentiate PC isomers. This task is only possible by using additional information, such as chromatographic behavior and manual analysis of the relative intensity of fragments in the MS/MS spectra. Therefore, the combination of initial screening by using MS-FINDER and SIRIUS with manual analyses of additional information is a powerful and efficient approach for identifying phenolic compounds.

Keywords: bioinformatics; high-resolution mass spectrometry; identification; phenolic compounds.