Supplementary Figure 9: DIA analysis using predicted spectral libraries.

(a) Impact of the retention time refinement using Prosit on the number identified peptides using either the general (indicated by “-”) or refined (indicated by “+”) Prosit indexed retention time (iRT) prediction model. The number of shared (blue), gained (green) and lost (orange) identified peptide sequences is plotted with respect to the original filtered library. iRT refinement was performed using the experimental retention time of the filtered HEK-293 data. See Supplementary Figure 8 for iRT model refinement analysis. (b) Identical analysis as Figure 4 for S. cerevisiae and E. coli. (c) Re-analysis of Orbitrap/TOF based data independent acquisition (DIA)/SWATH datasets using predicted spectral libraries. Data and project specific spectral libraries were obtained from public repositories. To facilitate comparisons, the original library was filtered for entries that Prosit is not yet able to predict (other modifications besides oxidized methionine, neutral losses and peptides >30 amino acids). The original and filtered spectral libraries were queried against the DIA data using Spectronaut and the barcharts depict the number of shared (blue), gained (green) and lost (orange) identified peptide sequences when using the original filtered library compared to the original unfiltered library. (d) Identical analysis as Figure 4, however protein-groups instead of peptides are displayed