From mass spectral features to molecules in molecular networks: MolNotator LDB dataset.

Finding actual molecules in LC-MS/MS experiments can prove challenging due to the considerable amount of redundant ions generated during ionization. In this context, MolNotator was created and validation with this dataset. MolNotator is a Python 3.7 package designed to predict molecules (molecular masses) by combinatorial triangulation in LC-MS/MS experiments after a preprocessing step using MZMine. An MGF and a CSV files output from MZmine are required as input for MolNotator which are placed in the "mzmine_out" folder of the project folder (the uploaded dataset). Instructions for the use of MolNotator are available on GitHub (https://github.com/ZzakB/MolNotator), Pypi (https://pypi.org/project/MolNotator/) and in the associated publication. The dataset consists of 193 LC-MS/MS analyses of lichen pure standards (previously used for the Lichen Database, LDB), 156 of which were detected by manual curation and served to benchmark MolNotator. Results indicated more than 90% of the 156 molecular masses were predicted by MolNotator under 2 ppm error on average. The project folder contains, in addition to the mzmine_out folder, a database and a params folder (see GitHub and Pypi) as well as a styles folder, containing different styles that can be imported on Cytoscape (https://cytoscape.org/) to visualise MolNotator's network output.