Traditional natural products discovery workflows implying a combination of different targeting strategies including structure- and/or bioactivity-based approaches, afford no information about new compound structure until late in the discovery pipeline. By integrating a MS/MS prediction module and a collaborative library of (bio)chemical transformations, we have developed a new platform, coined MetWork, that is able of anticipating the structural identity of metabolites starting from any identified compound. In our quest to discover new monoterpene indole alkaloids, we demonstrate the utility of the MetWork platform by anticipating the structures of five previously undescribed sarpagine-like N-oxide alkaloids that have been targeted and isolated from the leaves of Alstonia balansae using a molecular networking-based dereplication strategy fueled by computer-generated annotations. This study constitutes the first example of a natural product discovery workflow, termed CANPA, in which the targeted structures were initially generated and therefore anticipated by a computer prior to their isolation.
Mehdi Beniddir Université Paris-Sud-Faculty of pharmacy
download asset manuscript.pdf 1 MB [opens in a new tab] cloud_download
pdf : 1 MB
download asset SI.docx 14 MB [opens in a new tab] cloud_download
docx : 14 MB