Public code, models, and datasets for MS2Quant2 which is a machine learning tool to predict the ionisation efficiency of identified and unidentified chemicals in water by nontarget LC/HRMS
Here we will publish the code, models, and datasets of MS2Quant2, which is the updated version of MS2Quant.
The new version benefits from larger datasets and supports both negative and positive ionization modes.
MS2Quant2 enables prediction of ionization efficiencies directly from MS/MS spectra (without compound identification) via SIRIUS fingerprints, as well as from molecular structures. It facilitate quantification in targeted and non‑targeted analysis.
Predictions will be available from Python and R, alongside several new features to improve usability and performance.
For the initial version of MS2Quant, see:
The updated version of MS2Quant has been made possible thanks to dataset contributions from the community.
We gratefully acknowledge contributors:
Positive ion mode dataset: Emma Apelgren, Sam Mauritz Badii, Riccardo Costalunga, Marius Gaedke, Artur Gornischeff, Julianne Hollender, Carolin Huber, Lisa Jonsson, Karin Kiefer, Gunda Köllensperger and team, Jaanus Liigand, Louise Malm and NORMAN network, Miklos Mohai, Merit Oss, Irene Pulido Campillo, Riin Rebane, Helen Sepman, Jon Sobus, Tingting Wang, Wei-Chieh Wang, Anton Wällstedt.
Negative ion mode dataset: Lena Aisch, Nahid Amini, Emma Apelgren, Sam Mauritz Badii, Marius Gaedke, Juliane Hollender, Carolin Huber, Lisa Jonsson, Sara Khabazbashi, Gunda Köllensperger and team, Anneli Kruve, Mélanie Z. Lauria, Thomas Ledbetter, Jaanus Liigand, Piia Liigand, Louise Malm, Corina Meyer, Mari Ojakivi, Irene Pulido Campillo, Helen Sepman, Jon Sobus and team, Anton Wällstedt, Wei-Chieh Wang.
If you can provide additional ionisation efficiency datasets, please contact Anneli Kruve.