in silico Automatic Annotation of Phenolamides in Plants by Tandem Mass Spectra

Phenolamides (PAs) exist extensively in plants. They are comprised of the conjugation of a phenolic structure with polyamines or deaminated aromatic amino acids. They participate in multiple physiological process in plants, e.g. organ growth, adaptation and defense. However, this process is hindered by the existence of many different isomers of phenolamides and the lack of reference standards. A big problem is that it is time-consuming to process the data which leads to the identification of these compounds being incredibly inefficient. The purpose of this study is to establish a novel strategy to automatically annotate phenolamides by a machine learning model. The various typical phenolamides were synthesized according to the structural diversity of their building blocks. Liquid chromatography – mass spectrometry (LC-MS) was used to summarize the fragmentation patterns. A fragmentation pattern-based identification strategy was utilized and a machine learning model was developed by using Python script in order to identify hundreds of phenolamides in hours. The developed method was applied to identify phenolamides in herbal medicines and vegetables. These results will be useful for medicinal purposes.