Chemoinformatic Tools Used in Untargeted Metabolomic Approaches by Mass Spectrometry Applied for Natural Product Analyses

The chemistry of natural products has undergone a major transformation in the last twenty years, largely due to the development of powerful coupling techniques such as LC-HRMS/MS. These techniques, combined with supervised and unsupervised multivariate statistical analyses, are used for untargeted metabolomic studies for a wide range of applications. They have also enabled the development of dereplication approaches, thus accelerating often lengthy purification processes by focusing on biologically active metabolites of unknown structure. These dereplication approaches have been further strengthened in recent years by the development of molecular networks, based on the principle of grouping compounds according to their fragmentation profile in mass spectrometry. One of the current challenges remains the annotation of a large number of variables with a high degree of confidence. This will require enriching existing databases, and more recently, leveraging artificial intelligence. The latter, integrating in-silico virtual screening and chemoinformatic approaches, is now emerging as a powerful tool for predicting biological activity.