Fragment ion intensity prediction improves the identification rate of non-tryptic peptides in timsTOF

Abstract

Immunopeptidomics plays a crucial role in identifying targets for immunotherapy and vaccine development. Because the generation of immunopeptides from their parent proteins does not adhere to clear-cut rules, rather than being able to use known digestion patterns, every possible protein subsequence within HLA class-specific length restrictions needs to be considered. This leads to an inflation of the search space and results in lower spectrum annotation rates. Rescoring is a powerful enhancement of standard sequence database searching that boosts the spectrum annotation performance. In the field of immunopeptidomics low abundant peptides often occur, which is why the highly sensitive timsTOF instruments are increasingly gaining popularity. To improve rescoring for immunopeptides measured using timsTOF instruments, we trained a deep learning-based fragment ion intensity prediction model. Over 300,000 synthesized non-tryptic peptides from the ProteomeTools project were analyzed on a timsTOF-Pro to generate a dataset that was used to fine-tune an existing Prosit model. By applying our fragment ion intensity prediction model, we demonstrate up to 3-fold improvement in the identification of immunopeptides. Furthermore, our approach increased detection of immunopeptides even from low input samples.

A complete mass spectrometry-based immunopeptidomics pipeline for neoantigen identification and validation

“…Application of DeepNovo Peptidome to expand the reference HLA Ligand Atlas of benign human tissues As neoantigens mostly arise from non-canonical sources which are challenging for standard database search approaches, we built DeepNovo Peptidome, a de novo sequencing-based workflow for HLA peptide identification and neoantigen discovery. We trained the de novo sequencing model 13,14 on a very large, carefully curated dataset from 20 previous MS-based immunopeptidomics studies 5,15,16,[24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40] . The dataset included nearly 3,000 runs on two major MS instruments Orbitrap and timsTOF, >26M PSMs and >1.1M unique HLA peptides, and covering more than 90 alleles for each HLA class.…”

Emerging potential of immunopeptidomics by mass spectrometry in cancer immunotherapy

“…In addition to the previously used metrics, mass to charge ration (m/z), retention time (RT) and IM, the additional metrics, such as predicted spectral angle (SA) and the predicted RT, have been introduced to enhance the (immune)peptide identification. 57 , 58 , 59 These synergic progresses both in MS hardware and the supportive search systems make robust identification possible from the growing large‐scale MS raw data. As handling of gigantic MS data has become more generalized, machine‐learned search‐assisting algorithms play a greater role in this field.…”

TIMS2Rescore: A DDA-PASEF optimized data-driven rescoring pipeline based on MS2Rescore

“…We, and others, have previously shown that different fragmentation methods can heavily alter peptide MS2 spectra (27,28). We therefore trained new MS²PIP models that can accurately predict peak intensities for timsTOF acquired peptides.…”