Identification of modified peptides using localization-aware open search

Fengchao Yu,Alexey I Nesvizhskii,Andy T Kong,Dmitry M Avtonomov,Guo Ci Teo,Daniel J Geiszler,Sarah E Haynes

doi:10.1038/s41467-020-17921-y

Fengchao Yu, Alexey I Nesvizhskii + Show 5 more

Open Access

https://doi.org/10.1038/s41467-020-17921-y

Copy DOI

Journal: Nature Communications	Publication Date: Aug 13, 2020
Citations: 146	License type: open-access

Affiliation: University of Michigan–Ann Arbor

Abstract

Identification of post-translationally or chemically modified peptides in mass spectrometry-based proteomics experiments is a crucial yet challenging task. We have recently introduced a fragment ion indexing method and the MSFragger search engine to empower an open search strategy for comprehensive analysis of modified peptides. However, this strategy does not consider fragment ions shifted by unknown modifications, preventing modification localization and limiting the sensitivity of the search. Here we present a localization-aware open search method, in which both modification-containing (shifted) and regular fragment ions are indexed and used in scoring. We also implement a fast mass calibration and optimization method, allowing optimization of the mass tolerances and other key search parameters. We demonstrate that MSFragger with mass calibration and localization-aware open search identifies modified peptides with significantly higher sensitivity and accuracy. Comparing MSFragger to other modification-focused tools (pFind3, MetaMorpheus, and TagGraph) shows that MSFragger remains an excellent option for fast, comprehensive, and sensitive searches for modified peptides in shotgun proteomics data.

Highlights

Identification of post-translationally or chemically modified peptides in mass spectrometrybased proteomics experiments is a crucial yet challenging task
This advance enabled a practical implementation of the open search strategy, where large mass differences are allowed between unmodified peptide sequences and experimentally observed precursors
Fragment ions containing unknown modifications are not used in peak matching and scoring, and only half of all theoretical fragment ions can be matched to an experimental spectrum

Summary

Introduction

Identification of post-translationally or chemically modified peptides in mass spectrometrybased proteomics experiments is a crucial yet challenging task. We have recently introduced a fragment ion indexing method and the MSFragger search engine to empower an open search strategy for comprehensive analysis of modified peptides. MS-based proteomic datasets contain a large number of peptides harboring chemical modifications introduced at the sample preparation stage[4,5] Searching for these many possible modifications using conventional database search tools has proven impractical. We introduced a fragment ion indexing strategy and an ultrafast database search engine MSFragger[15] This advance enabled a practical implementation of the open search strategy, where large mass differences are allowed between unmodified peptide sequences and experimentally observed precursors. Applications have included detection of RNA crosslinks on RNA binding proteins[16], comprehensive PTM searching in metaproteomics studies[17] and clinical samples[18], quality control and checks for unexpected modifications[19], and proteogenomics[20]

Methods

Results

Conclusion