Improved open modification searching via unified spectral search with predicted libraries and enhanced vector representations in ANN-SoLo

The iPRG2012 dataset, created by the ABRF Proteome Informatics Research Group, is a challenging benchmark for evaluating post-translational modification (PTM) identification methods. It contains 17,993 MS/MS spectra from synthetic PTM peptides spiked into a yeast lysate background. To evaluate the new features of ANN-SoLo, three spectral libraries were used: Splib: A combined reference library from yeast (TripleTOF) and human (NIST HCD) spectra, processed with SpectraST to generate decoys, totaling 1,180,014 spectra. Decoy Internal: Contains the same target spectra as Splib, but with internally generated decoys, totaling 1,172,923 spectra. Prosit: A large predicted library (14,672,189 spectra) generated from the iPRG2012 FASTA file using Prosit, with tryptic peptides (using 2 missed cleavages) on two charges: 2 & 3. This setup enables benchmarking ANN-SoLo against a reference, an internally generated decoys, and predicted spectral library. The repository includes: decoy_internal_library.rar, prosit_library.rar, splib_library.rar: Spectral libraries decoy_internal.mztab, prosit.mztab, splib.mztab: Corresponding identification results ABRF_iPRG_2012_target.fasta: Sequence database iPRG2012.mgf: Query spectra file