Extending compound identification for molecular network using the LipidXplorer database independent method: A proof of concept using glycoalkaloids from Solanum pseudoquina A. St.-Hil.

INTRODUCTION: Molecular networks are now established as the method of choice for tandem mass spectrometry dereplication and similarity-based structure elucidation. Node identification can be used to start the propagation of the structure elucidation of unknown compounds progressively.

OBJECTIVE: To demonstrate the capabilities of using the LipidXplorer data results along with molecular networking to identify nodes and aid sequential structure elucidation of unknown compounds.

MATERIAL AND METHODS: Molecular fragmentation query language (MFQL) files were written to identify glycoalkaloids based on known structures described for Solanum species. A dataset generated from liquid chromatography-high resolution mass spectrometry (LC-HRMS) analysis of Solanum pseudoquina sample were submitted to dereplication on both LipidXplorer software and Global Natural Products Social Molecular Network (GNPS) online system. The resulting attribute table from GNPS calculations was merged with the LipidXplorer results and this merged file was used for network visualisation in Cytoscape. Nodes in the molecular network were labelled using the LipidXplorer identifiers, thus assisting the structure elucidation of unidentified compounds.

RESULTS: The combination of the LipidXplorer glycoalkaloids list and GNPS analysis was used in Cytoscape to label nodes in the molecular network. The analysis of the network using these labelled starting points triggered the structure elucidation of closely related nodes leading to the identification of 30 compounds using the LipidXplorer output and four purified and structure elucidated compounds, including a new glycoalkaloids identified as 3-O-(β-d-xylopyranosyl)-(20R,25S)-22,26-epimino-16-acetyl-cholesta-5,22(N)-diene.

CONCLUSION: A significant compound identification completely based on molecular formula and fragmentation queries was achieved. This new and effective approach could help researches to expand the identification rate of compounds in dereplication studies using molecular networks.