An in-house database-driven untargeted identification strategy for deep profiling of chemicalome in Chinese medicinal formula

Deep profiling of chemicalome in Chinese medicinal formulas is vital for disclosing the secret underlying their effectiveness. To address this issue, an in-house database-driven untargeted identification strategy was proposed with the use of ultra-performance liquid chromatography coupled to quadrupole time of flight mass spectrometry. Firstly, an in-house mass spectral database for the analyzed herbs was constructed, and database querying was performed for rapid recognition of known compounds. Secondly, a chemical diagnostic characteristics algorithm was originally developed for deep mining unrecorded ions, and thus expanding coverage of components beyond the database. Additionally, we proposed evaluation criteria for the untargeted identification of compounds with different confidence levels. As a case study, the integrated strategy was applied to comprehensively characterize complex multi-type components in Gegen-Qinlian Decoction. A total of 381 compounds were characterized and annotated with four different confidence levels, and 88.40% of these annotated compounds were successfully re-identified in triplicate analyses with a different instrument. The integrated strategy was demonstrated powerful in deep profiling of chemicalome in Chinese medicinal formulas with higher throughput, analytical sharpness, and lower omission ratios.