INFERYS
RESCORING

Unlocking the intensity dimension

Database search engines for bottom-up proteomics largely ignore peptide fragment ion intensities during the automated scoring of tandem mass spectra against protein databases. Using predicted fragment ion intensities to calculate additional scores helps to overcome this drawback. INFERYS™ Rescoring for Sequest™ HT is an integrated workflow in Thermo Scientific™ Proteome Discoverer™ software, promising a deeper, more confident and comprehensive analysis of proteomic data.

Left panel: schematic workflow of INFERYS Rescoring. Briefly, all fragment ion intensites of all peptide-spectrum matches (PSMs) identified by Sequest HT are predicted and used for intensity-based scoring. These additional scores are combined with classical scores and submitted to Percolator for score aggregation and false discovery rate (FDR) estimation.

Rescoring boosts peptide identifications at 1% FDR

Rescoring fuses the well-established concept of database search with the advantages of state-of-the-art deep learning on your local PC. Rescoring provides merit for all sample types but demonstrates particular strength in challenging samples like immunopeptidomics experiments. Here, Rescoring enables better separating target and decoy identifications, thereby boosting the number of identified PSMs and peptides, while retaining the same false discovery rate.

Top left: number of PSMs and peptides identified by a conventional workflow and INFERYS Rescoring for an immunopeptidomic HLA Class I dataset from a melanoma cell line (Chong et al. Nat. Com, 2020). Bottom left: Percolator SVM score distribution for the conventional and the INFERYS Rescoring workflow demonstrate better separation of target and decoy identifications, leading to a higher identification rate and better confidence. Bottom right: motif analysis for the HLA-A*02:04 allele comparing the identified peptides from both workflows and INFERYS Rescoring-exlcusive identifications, resulting in a similar 9mer motif.

INFERYS Rescoring is integrated into Proteome Discoverer software. For more information and licensing, please visit thermofisher.com/proteomediscoverer. If you would like to learn what a customized deep learning model can do for your research, feel free to contact us!

Thermo Scientific™ Proteome Discoverer™ software is a trademark of Thermo Fisher Scientific, Inc. SEQUEST is a trademark of the University of Washington.