Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry.
Collins BC, Hunter CL, Liu Y, Schilling B, Rosenberger G, Bader SL, Chan DW, Gibson BW, Gingras AC, Held JM, Hirayama-Kurogi M, Hou G, Krisp C, Larsen B, Lin L, Liu S, Molloy MP, Moritz RL, Ohtsuki S, Schlapbach R, Selevsek N, Thomas SN, Tzeng SC, Zhang H, Aebersold R. · 2017
Read Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry. by Collins BC, Hunter CL, Liu Y, Schilling B, Rosenberger G, Bader SL, Chan DW, Gibson BW, Gingras AC, Held JM, Hirayama-Kurogi M, Hou G, Krisp C, Larsen B, Lin L, Liu S, Molloy MP, Moritz RL, Ohtsuki S, Schlapbach R, Selevsek N, Thomas SN, Tzeng SC, Zhang H, Aebersold R. (2017) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc5566333.txt.
Introduction
Collins and co-authors report a multi-laboratory test of SWATH mass spectrometry in Nature Communications. The method aims to count proteins at large scale with targeted accuracy. Single labs had shown steady results. Whether labs agree with each other was untested.
Eleven sites worldwide ran the same samples. All could detect and count over 4,000 proteins from HEK293 cells. A synthetic peptide dilution series checked sensitivity and range. Those traits came out uniform across sites. The study gives the field a shared benchmark. It builds trust that SWATH data can be reproduced between labs. Readers see rare, direct proof of an omics method traveling well.
Catalog record
| CREATOR | Collins BC, Hunter CL, Liu Y, Schilling B, Rosenberger G, Bader SL, Chan DW, Gibson BW, Gingras AC, Held JM, Hirayama-Kurogi M, Hou G, Krisp C, Larsen B, Lin L, Liu S, Molloy MP, Moritz RL, Ohtsuki S, Schlapbach R, Selevsek N, Thomas SN, Tzeng SC, Zhang H, Aebersold R. |
|---|---|
| YEAR | 2017 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 73 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry By Ben C. Collins, Christie L. Hunter, Yansheng Liu, Birgit Schilling, George Rosenberger, Samuel L. Bader, Daniel W. Chan, Bradford W. Gibson, Anne-Claude Gingras, Jason M. Held, Mio Hirayama-Kurogi, Guixue Hou Published in: Nature Communications DOI: 10.1038/s41467-017-00249-5 ABSTRACT Quantitative proteomics employing mass spectrometry is an indispensable tool in life science research. Targeted proteomics has emerged as a powerful approach for reproducible quantification but is limited in the number of proteins quantified. SWATH-mass spectrometry consists of data-independent acquisition and a targeted data analysis strategy that aims to maintain the favorable quantitative characteristics (accuracy, sensitivity, and selectivity) of targeted proteomics at large scale. While previous SWATH-mass spectrometry studies have shown high intra-lab reproducibility, this has not been evaluated between labs. In this multi-laboratory evaluation study including 11 sites worldwide, we demonstrate that using SWATH-mass spectrometry data acquisition we can consistently detect and reproducibly quantify >4000 proteins from HEK293 cells. Using synthetic peptide dilution series, we show that the sensitivity, dynamic range and reproducibility established with SWATH-mass spectrometry are uniformly achieved. This study demonstrates that the acquisition of reproducible quantitative proteomics data by multiple labs is achievable, and broadly serves to increase confidence in SWATH-mass spectrometry data acquisition as a reproducible method for large-scale protein quantification. INTRODUCTION Reproducibility is an essential foundation of scientific research. Recent reports have concluded that a significant fraction of life science research shows poor reproducibility of results and this poses a major challenge to scientists, science policy makers, funding agencies, and the pharma and biotech industry sectors1–3. The reasons for irreproducibility of research results are many, including inadequate study design and data analysis, limited data quality, incompletely characterized research reagents, poorly benchmarked techniques, and a range of other confounding factors. The question of whether specific data acquisition methods and platforms are capable of generating reproducible results is best addressed by inter-laboratory studies, where samples of known composition and quality are analyzed across different settings. Such studies have been reported for various “omics” technologies, including RNA-seq and microarray techniques, with varying results4, 5. Such projects have served to highlight problems in various large-scale strategies, to stimulate discussion in a given field on how to improve reproducibility, and in the best cases to provide confidence in a given strategy within and beyond an analytical field. In the field of mass spectrometry (MS) based proteomics, a wide range of specific methods have been reported over the past two decades. These can be broadly grouped into discovery and targeted proteomic techniques. The general aim of discovery proteomics is the unbiased identification and quantification of the protein components of biological samples. This is most frequently achieved by data-dependent acquisition (DDA) MS. If the number of precursor ions exceeds the number of precursor selection cycles6, precursor selection becomes stochastic and the peptides detected in repeat analyses become irreproducible. This has been documented in a number of intra- and inter-laboratory studies7–9. In general, these studies confirmed that a high degree of reproducibility is difficult to achieve for complex samples10. Computational methods to enable improved quantification via propagation of peptide identifications across runs via alignment of MS1 precursor signals, first introduced as accurate mass and time tags (AMT)11, 12, are commonly applied to DDA data13–16 and can reduce this issue to some degree in discrete data sets where chromatographic alignment can reasonably be applied. In contrast to discovery proteomics the general aim of targeted proteomics is the detection and quantification of a predetermined set of peptides by selected reaction monitoring (SRM) also known as multiple reaction monitoring (MRM)17, or a related technique parallel reaction monitoring18–20. Because targeted MS eliminates the stochastic component of precursor ion selection in DDA, it has the potential for high reproducibility. This has been demonstrated in intra-laboratory studies where sets of peptides were targeted with a high degree of reproducibility across relatively large sample sets21–23 and by inter-laboratory studies focused on exploring the use of SRM and immuno-SRM for biomarker studies24–29. Targeted MS is now broadly regarded as a reproducible protein analysis platform17. However, the number of proteins measured is restricted (usually to ~100 per injection), limiting its utility for many applications. SWATH-MS is a more recently introduced approach to MS-based proteomics30. It consists of data-independent acquisition31, 32 (DIA) in which all precursor ions within a user defined m/z window are deterministically fragmented. Analysis of SWATH-MS data most often relies on a targeted data analysis strategy in which target peptides are detected and quantified from the SWATH-MS fragmentation data by extracting and correlating previously generated query parameters for each target. In this scheme each unique peptide of interest at a given precursor charge state is queried for in the data, resulting in the detection and scoring of co-eluting transition group signals and associated underlying mass spectral features, referred to as peak groups. Because the method specifically tests for the presence of each target peptide in the essentially complete fragment ion map of each sample, it eliminates the stochastic sampling element of DDA and helpfully provides a direct statistical measure (e.g., q-value) of whether the peptide is present at a detectable level in the sample. This data analysis strategy, whereby target peptides are directly queried for, has recently been generalized using the term peptide-centric33, 34 scoring to distinguish from more classical approaches where the MS2 spectrum is the query unit for data analysis (referred to as spectrum-centric scoring). The SWATH-MS implementation of the DIA concept therefore preserves the favorable performance characteristics of SRM, while vastly expanding the measurement capacity to thousands of proteins per injection. Of consideration in SWATH-MS is the complexity of the resultant spectra and specific software tools have been compiled to analyze such highly multiplexed data using various approaches35–38. A recent study comparing software tools for the analysis of DIA data using either peptide-centric or spectrum-centric approaches has demonstrated that very similar qualitative and quantitative results can be obtained when analyzing a benchmarking data set39. SWATH-MS and related DIA approaches have achieved a high degree of reproducibility in intra-laboratory studies in a variety of research questions such as interaction proteomics40, 41, plasma proteomics42, tissue proteomics43, microbial proteomics44, 45, pre-clinical toxicology9, analysis of genetic reference strains46, and many others. However, interlaboratory robustness and reproducibility of SWATH-MS data acquisition has not been demonstrated. In this study, we set out to test the reproducibility of peptide and inferred protein detection and quantification by SWATH-MS in an inter-laboratory study. To achieve this goal we distributed benchmarking samples to 11 participating laboratories worldwide for measurement by SWATH-MS according to a predetermined schedule. We analyzed the data from all sites centrally with two separate scopes in mind. Firstly, we analyzed all of the data in an aggregated way to simulate, for example, a large cohort study whereby patient samples would be analyzed in multiple laboratories, aiming to achieve a result set based on all samples. In the second interpretation, we analyzed the data from each site of collection independently and compared the results across sites post analysis facilitating a direct performance comparison. Our analysis demonstrated that the set of proteins detected and quantified across all participating sites, i.e., from a total of 229 proteome measurements, was very consistent. The reproducibility, linear dynamic range, and sensitivity