Positive-Unlabeled Learning for Predicting Small Molecule MS2 Identifiability from MS1 Context and Acquisition Parameters
Madina Bekbergenova, Tao Jiang, Louis-Félix Nothias-Scaglia, Wout Bittremieux
The paper develops a deep learning model to predict MS2 identifiability using MS1 data and acquisition parameters, achieving 90% recovery of known identifiable spectra.
bioRxiv · Jan 24, 2026 · PreprintBioinformatics