In this study, we extended the ProteomeTools peptide library (PROPEL, see PXD004732 and PXD010595) to train a deep neural network resulting in chromatographic retention time and fragment ion intensity predictions for (tryptic) peptides that exceed the quality of the experimental data.