Fig. 1: Data collection and collision energy calibration for refining Prosit to Prosit-XL.

a Pie chart showing the collected training data on cross-linked spectrum match (CSM) and unique XL-peptide (peptide pair) level, covering the cleavable cross-linkers Thiol and Alkene acquired using MS3 spectra (CMS3) and DSSO and DSBU acquired using MS2 spectra (CMS2), as well as data for non-cleavable cross-linkers DSS/BS3 acquired using MS2 spectra (NMS2). b Normalized collision energy (NCE) calibration curve for an example MS-file, showing the mean spectral angle (SA) when comparing annotated experimentally acquired spectra of the top 1000 highest scoring target peptide-spectrum matches (PSMs) to spectra predicted with Prosit at varying NCEs. The NCE with the highest average SA, indicated by the vertical red line, is used as the NCE for training. c, d Bar plots showing the number of raw files after NCEs calibration across the training data. e Violin plot comparing the annotated experimental MS2-MS3 spectra of XL-peptides with the same peptide A but different peptide B for five different cross-linkers: CMS3-Thiol, CMS3-Alkene, CMS2-DSBU, NMS2-DSS/BS3, and CMS2-DSSO. The analysis is not focused on a specific peptide A; instead, peptide A refers to the first peptide in each cross-linked pair. The number of sampled spectra (n = 1700) is indicated at the bottom. The black solid line and corresponding numbers indicate the median spectral angle (SA) and Pearson correlation (PCC) for each distribution. Mean spectral angles ± standard error of the mean (SEM) for each group are as follows: Thiol(CMS3), 0.857 ± 0.003; Alkene(CMS3), 0.827 ± 0.003; DSBU(CMS2), 0.779 ± 0.005; DSS/BS3(NMS2), 0.785 ± 0.003; and DSSO(CMS2), 0.724 ± 0.004.