Additional Data from: Machine Learning for Enhanced Identification Probability in RPLC/HRMS Non-Targeted Workflows

§  Additional file 1 A spreadsheet with a list of compounds used to train the pre-trained model, the molecular fingerprint-to-retention time index (MF-to-RTI) model. §  Additional file 2 A spreadsheet with a list of pre-selected cumulative neutral loss (CNL) masses-of-interest and their corresponding quantities. §  Additional file 3 A spreadsheet with a list of compounds and their corresponding quantities of tandem mass spectra used for training the cumulative neutral loss-to-retention time index (CNL-to-RTI) model. §  Additional file 4 A spreadsheet file with 3 sheets with a list of true RTI values of compound calibrants used in Model 1’s training and testing with publicly available MS/MS spectra, accompanying the predicted values from Models 1 and 2, and the RTI error from their true values for training (sheet 1), testing (sheet 2), and assessing the calibrants with RTIs from different systems (sheet 3). §  Additional file 5 A spreadsheet file with 4 sheets collecting all reference MS/MS spectral matching results for the compound Pencycuron in various degrees of black tea matrix (sheet 1, summary; sheet 2, data for samples with no tea matrix; sheet 3, data for samples with the 100x-diluted matrix; sheet 3, data for samples with the 10x-diluted matrix).