Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2025 Oct 22;24(11):5904–5910. doi: 10.1021/acs.jproteome.5c00637

Considering Measurement Time and Depth in diaPASEF Plasma Proteomics

Eva R Smit 1, Carmen van der Zwaan 1, Stijn A Groten 1, Maartje van den Biggelaar 1, Arie J Hoogendijk 1, Pieter Langerhorst 1,*
PMCID: PMC12604034  PMID: 41125229

Abstract

Mass spectrometry (MS)-based plasma proteomics is a powerful approach to unraveling the biology or pathophysiology in large clinical cohorts. Implementation of data-independent acquisition in combination with ion mobility approaches (diaPASEF) has enabled high-throughput analysis with increasing proteomic depth. DIA-based methods are dependent upon experimental or in-silico-generated spectral libraries, yet there is a lack of consensus in the field about which library produces superior results in regard to protein and peptide identifications and quantifications. Here, we evaluated approaches for building a spectral library in plasma proteomics on a timsTOF HT system. Furthermore, the relationship between measurement time, library depth, and number of protein and peptide identification for high-throughput plasma proteomics applications was assessed. As expected, an increase in the measurement time invested in the spectral library enhanced the number of identifications. At the protein level, in silico libraries provided decreased depth compared to the most extensive experimental library. However, the experimental library enhanced the number of peptide identifications by 14% compared to that of the in silico library. With the field increasingly moving to peptide-centric approaches, an experimental library allows for deeper assessment in peptide-based studies.

graphic file with name pr5c00637_0004.jpg

graphic file with name pr5c00637_0003.jpg

Introduction

Unbiased plasma proteomics is a powerful technique to unravel the biology, disease pathophysiology, or new biomarkers through studying blood circulating proteins. − Mass spectrometry (MS)-based proteomics allows for the unbiased identification and quantification of plasma proteins with amino acid level specificity. However, plasma has proven a challenging matrix due its large dynamic range of protein abundances, − which limits the detection of low-abundance proteins, e.g. tissue-leakage proteins, chemokines, and other signaling proteins. ,

To improve plasma proteomic depth, several strategies have been developed, focusing either on sample treatment or MS-data acquisition. Sample treatment generally aims at decreasing dynamic range, − but results in an increased cost per sample. Alternatively, improved MS-data acquisition methods have been established, i.e., independent acquisition (DIA) that offers both an increased proteomic depth and higher quantification accuracy. Additionally, DIA can be used in combination with relatively short liquid chromatography (LC) gradients while maintaining proteomic depth for high-throughput applications. − Recently, the addition of ion mobility (IM) mass spectrometric analysis, used on timsTOF instruments with parallel accumulation serial fragmentation (diaPASEF), is increasingly used. This additional dimension has been shown to separate isobaric peptides and increase scanning speeds. − These developments further enhanced the application of plasma proteomics in clinical proteomic studies. ,

Notably, analysis of DIA data dependents on experimentally obtained spectral libraries − or in-silico-generated libraries (DIA-NN, AlphaPeptDeep), , henceforward in silico libraries. The wide variety of library creation methods in plasma proteomic studies indicates a lack of consensus in the field, especially for the next-generation IM-MS techniques. ,− To this end, a previous study has already shown that the optimization of LC gradients and diaPASEF windows is beneficial for the identification of proteins, which can be further boosted by the inclusion of extracellular vesicles. However, the effects that experimental spectral libraries have on both the proteomic depth and peptide depth have not yet been evaluated. Here, we systematically investigated the impact and implications of library creation strategies in neat plasma on protein and peptide identification with a timsTOF HT system.

Experimental Section

Plasma Isolation

Citrated whole blood was drawn from 10 healthy volunteers. Plasma was isolated from whole blood by centrifugation for 10 min at 2000g at room temperature to collect plasma after which the plasma was again centrifugated for 10 min at 2000g at room temperature. Aliquots of plasma were stored at −80 °C until the experiment. Ethical approval was obtained from the Sanquin Ethical Advisory Board.

Plasma Sample Preparation

Plasma samples were thawed at 37 °C and 10 μL was diluted in 590 μL of 100 mM Tris­(hydroxymethyl)­aminomethane hydrochloride (Tris-HCl, Life Technologies, U.K.) (pH 8.0). Nine μL of 60× diluted plasma was resuspended in 5 μL of reduction and alkylation buffer (20 mM Tris­(2-carboxyethyl)­phosphine (TCEP, Thermo Fisher Scientific, Rockford, IL), 80 mM chloroacetamide (CAA) (Sigma–Aldrich, St Louis, MO) in 100 mM Tris-HCl (pH 8.0)) and then heated at 95 °C for 5 min. After the samples were cooled to room temperature 30 μL of 50 mM Tris-HCl (pH 8.0), containing 200 ng (1:50 ratio) MS-grade Trypsin Gold (Promega, Madison, WI) was added and samples were continuously shaken at 700 rpm at 25 °C overnight for digestion. The digestion was quenched with 5 μL of 10% trifluoroacetic acid (TFA, Thermo Fisher Scientific, Rockford, IL) and kept at −20 °C until analysis.

Liquid Chromatography

For proteomics analysis, 500 ng of peptide mixture was loaded in triplicate onto Evotip PureTips (Evosep, Denmark) according to manufactures guidelines. Gradient was set to either 15, 30, or 60 SPD with the appropriate manufacturer recommended analytical columns (Performance column, 8 cm × 150 μm, 1.5 μm (EV1109) and 15 cm × 150 μm, 1.5 μm (EV1137)). For separation, 0.1% formic acid in water (Biosolve, The Netherlands) and 0.1% formic acid in acetonitrile (Biosolve, The Netherlands) were used as mobile phases A and B, respectively.

Experimental Library

For experimental spectral libraries, an in-house available pool of plasma from 30 healthy donors from Sanquin (Amsterdam, The Netherlands) was thawed at 37 °C and 60-fold diluted in 100 mM Tris-HCl (pH 8.0). For the spectral libraries, the plasma proteomics sample preparation workflow was scaled up by a factor of 15.

Off-Line Fractionation

For the off-line fractionated library, 1.2 mL of the digested plasma pool were fractionated using in-house prepared StageTips with three layers of Empore styrene divinylbenzene–reverse-phase sulfonate (SDB-RPS, Supelco, Bellefonte, PA) adapted from Kulak et al. Prior to sample loading StageTips, they were washed with 65 μL of acetonitrile (ACN, BioSolve, The Netherlands) and 65 μL of 0.2% TFA. After sample loading, the tips were twice washed with 65 μL of 0.2% TFA and fractionated with 50 μL of (1) 75 mM ammonium formate (AF, Thermo Fisher Scientific, Rockford, IL) in 30% ACN, (2) 90 mM AF in 35% ACN, (3) 100 mM AF and 0.5% FA in 40% ACN, (4) 117 mM AF and 0.5% FA in 45% ACN, (5) 133 mM AF and 0.5% FA in 55% ACN, (6) 150 mM AF and 0.5% FA in 60% ACN, (7) 160 mM AF and 0.5% FA in 68% ACN, and (8) 2× 30 μL of 5% ammonium hydroxide (Merck, Germany) in 80% ACN. For elution, samples were centrifuged at 200g for at least 5 min or until all buffer had eluted of the StageTips. After elution, samples were dried in a speedvac at 30 °C under vacuum. For each of the eight fractions collected in triplicates, samples were reconstituted in 0.1% FA in water (BioSolve, The Netherlands) and loaded onto seven Evotip PureTips (Evosep, Denmark), according to the manufacturer guidelines for LC-MSMS.

Mass Spectrometry Data Acquisition

Data was acquired on a timsTOF HT Mass Spectrometer (Bruker Daltonics, Billerica, MA) operated in ddaPASEF mode. Precursors were selected based on the default ion cloud filter, with a cycle time of 1.17 s, including 10 PASEF ramps. Accumulation time was set to 100 ms with a 100% duty cycle and target intensity of 20 000 (Threshold = 2500). Collision energy was set as a linear function of 1/k0 (0.6 1/k 0 = 20 eV; 1.60 1/k 0 = 59 eV). Ion mobility was set from 0.6 −1.6 1/k 0 in standard methods. For ion mobility fragmentation, the mobility ranges were set as follows: 0.6–0.8, 0.8–1.05, 0.9–1.15, 1–1.25, 1.2–1.45, and 1.35–1.6 1/k 0.

Data Independent Acquisition of Plasma Proteomes

MS-based plasma proteomics was performed on a timsTOF HT system operated in the diaPASEF mode. Ion mobility was set from 0.70 to 1.45 1/k 0, and collision energy, Ramp time, duty cycle as stated above. diaPASEF windows were established with pydiAID based on precursor density and chromatographic separation for 30/60SPD (). The cycle times were 3.07 and 1.85 s for the 30SPD and 60SPD method, respectively.

Data Processing

The spectral library was generated for each library with FragPipe (v23.0), using IonQuant (v1.11.9) and MSfragger (v4.2). , The reviewed human proteome database (Swiss-Prot Database, 20423 entries, downloaded on 8 August 8, 2023) was used as a database, including decoys (reverse) and common contaminants. The spectral library was generated with a peptide length of 6 to 50 amino acids was used with maximally 2 missed cleavages, a precursor, and fragment mass tolerance of 20 ppm, and FDR was set to 1% for PSM, peptide, and protein level.

DIA data was processed with DIA-NN (v2.0), using default settings based on the generated spectral library with a precursor FDR of 1%, a set mass and MS1 accuracy of 15 ppm, a precursor charge range from 1 to 4, a precursor m/z range from 300 to 1800, a minimum peptide length of 6 amino acids, a maximal peptide length of 30 amino acids. We allowed up to 2 missed cleavages and 3 variable modifications, allowing for methionine oxidation and protein N-terminal acetylation. We enabled match between runs, enabled protein inference, and used QuantUMS (high precision) as quantification strategy and peptidoform scoring. The same settings were used for the in-silico-predicted spectral library from the reviewed human proteome database (Swiss-Prot Database, 20423 entries, downloaded on August 8, 2023). All processing was performed on the same virtual machine.

To include the effects of restricted search space, we also processed 30SPD acquired DIA data with DIA-NN (v2.0), allowing up to 1 missed cleavage and 1 variable modification, with all other settings kept the same as previously described.

Data Analysis

All analysis and visualization was performed using python (v 3.10). To assess the content of spectral libraries was based on the library detected precursors in the MSfragger library.tsv for experimental data and the resultant library generated in DIA-NN for the in silico search. For analysis of proteins and precursors identified in DIA data, the pg_matrix.tsv and pr_matrix.tsv from DIA-NN were used, after filtering out common contaminants. The coefficients of variation (CVs) were calculated based on nonlog2 transformed label-free quantification (LFQ) values. Total measurement time was calculated as the sum of processing time extracted from log files and the experimental time determined by the total number of samples divided by the LC method.

Data Availability Statement

The mass spectrometry raw data (.d files) and processed data (.tsv) have been deposited in the ProteomeXchange/PRIDE archive database with identifier PXD058337.

Results and Discussion

To investigate the impact and influence of spectral library creations on the identification of proteins and peptides from neat plasma, we constructed spectral libraries by either off-line fractionation, IM fractionation, or combining these methods using a digested sample from an in-house available pool of plasma samples from 30 healthy donors from Sanquin (experimental libraries). The experimental libraries were acquired with ddaPASEF at different gradient lengths (15, 30, or 60 samples per day, SPD). Additionally we added an in-silico generated library (Figure A). First we sought to investigate the difference in library contents, utilizing the spectral library files generated from MS-data acquired from a single run. Library size for the highest yielding MS acquisition approach (off-line and IM) increased from 485 proteins to 727 proteins (50%) with increasing gradient length, without noticeable saturation. This was observed also for the other MS acquisition approaches (Figure B, ). Interestingly, the resulted in silico library after the DIA-NN search obtained with an in silico search of the entire reviewed proteome contained similar depth to the highest yielding experimental library, yet contained less precursors ().

1.

1

Open in a new tab

Experimental workflow and library specific results. (A) Schematic overview of the experimental design and spectral library building. (B) Comparison of library size based on proteins (top) and precursors (bottom). (C) Comparison of precursor charge distributions between libraries was based on the library detected precursors. (D) Comparison of fragment type distributions between libraries was based on the library detected precursors. (E) Comparison of fragment charge distributions between libraries based on the library detected precursors. (F) Precursor distribution displayed as kernel density estimation of precursors specific to the in silico build based on 30SPD plasma samples analyzed with diaPASEF. (G) Precursor distribution displayed as kernel density estimation of precursors shared between the in silico library build based on 30SPD plasma samples analyzed with diaPASEF and experimental library build using off-line and IM fractionation at 15SPD. (H) Precursor distribution displayed as kernel density estimation of precursors specific to the experimental library build using off-line and IM fractionation at 15SPD.

This prompted further investigation of the library composition. Although there was a high consistency for fragment type, fragment charge and precursors charge within experimental libraries, in silico libraries contained relatively more 1+/2+ charged precursors and less 3+/4+/5+ charged precursors (Figure C). Additionally, a higher contribution of y-ions and 1+ charged fragments was observed in in silico libraries (Figures D and E). We subsequently compared precursor distribution along the m/z and IM dimensions, stratified by shared and unique features of the in-silico and best performing experimental library, respectively (Figures F–H, ). The characteristic 2+ ion cloud is present in all three library types; however, the in silico library lacks the 4+/5+ precursor-ion cloud present in the experimental library (, Figure C). We and other have observed that plasma proteins have a higher percentage of >2+ charged precursors compared to cellular proteins. To our knowledge, research on algorithms for in-silico-generated libraries in diaPASEF have mostly utilized (cell) lysates to assess performance of these methods resulting in a potential bias toward cell-based, and thus 2+ charged, precursors. Therefore, we postulate that the discrepancy between the two approaches might be attributed to the inherent difference between plasma and cellular proteins.

To assess the impact of the libraries on protein and peptide identifications, citrated plasma samples from 10 healthy donors acquired with diaPASEF at 30 and 60SPD were benchmarked (Figure A). The smaller library, obtained from IM fractionation, resulted in the lowest library size and proteomic depth (388 proteins identified at 30SPD with 15SPD IM library), while combining off-line and IM fractionation with a 15SPD gradient resulted in the largest experimental library and highest proteomic depth in neat plasma (552 proteins identified). (Figures A and B). Overall, an increased library size resulted in enhanced proteomic depth with a supplemental effect of an increased gradient length. The gradient length has less influence on the proteomic depth than the library size. Next, we assessed the impact of the total time investment versus proteomic gains. The more time was allocated to obtaining the experimental library, enhanced the precursor and proteomic depth in neat plasma ranging from 2.8 experimental hours and 27.7 total analysis time with 286 proteins/5359 precursors identified to 78.4 experimental hours and 104.1 total analysis time with 552 proteins/9308 precursors identified, (Figures C and D). In contrast with previous research, we found that the in silico library yielded lower proteomic depth. The increased proteomic depth did not result in an overall lower mean data completeness or higher coefficients of variation ().

2.

2

Open in a new tab

Comparison of identified proteins and precursors by spectral library approach. (A) Dependency in proteins was identified from DIA data to the spectral library used. (B) Dependency in precursors identified from DIA data to the spectral library used. (C) Time investment in the spectral library building dependencies to proteins identified from DIA data. (D) Time investment in spectral library building dependency to precursors identified from DIA data. (E) Overlap in proteins identified at 30 SPD between the in silico library build and the experimental library, build using off-line and IM fractionation at 15SPD. (F) Log2 label-free intensity (LFQ) distribution for proteins shared between both spectral libraries (yellow) is unique to the in silico library (green) and the Experimental library (pink). (G) Overlap in precursors was identified at 30 SPD between the in silico library and experimental library. (H) Boxplot representing the distribution of unique peptides per identified protein in the in silico library (green) and experimental library (pink). (I) Von Willebrand Factor (VWF) protein coverage with peptides shared between both approaches (yellow) and unique peptides per identified protein in the in silico library (green) and the experimental library (pink).

Using the in silico library, we were able to identify less proteins in plasma samples than the best performing experimental library (462 in silico versus 552 experimental), indicating a suboptimal fitting of the in silico library data onto the plasma proteomic datasets. This can stem from the mismatch in library composition (Figures F–H). Yet, an in silico library forgoes the need of additional measurement time, resulting in advantageous protein per time investment ratio in comparison to the best performing experimental library (). However, the computational time per sample is higher when using an in silico library (a 12.2-fold increase, compared to experimental library), which might be a bottleneck in large clinical cohorts shifting the balance to an experimental approach.

Between the in silico and best performing experimental library approach, 433 proteins were identified in both, 27 uniquely in the in silico and 118 in the experimental library (Figure E, ). Out of the 27 uniquely identified proteins in the in silico library, 2 proteins (CDHR5, IGHV4–38–1) were also present in the experimental library. Yet the reason for the sole identification in the in silico library is unclear (). For the in silico library, the unique protein groups covered the entire LFQ range, whereas most unique protein groups for the experimental library are located toward to lower limit of LFQ, improving the dynamic range (Figure F). This could be especially relevant for applications in disease cohorts, as low abundant tissue-leakage proteins and disease-specific proteins can be expected to be present in those patients. No specific pathways nor networks could be identified from enrichment of spectral library specific proteins (data not shown). Second, we also compared the overlap in identified peptides between both library approaches and found that 5051 peptides were shared, although we again observed unique peptides to both approaches (Figure G). More peptides were detected with the experimental library, resulting in more peptides per proteins. For this non-normal distribution the interquartile range shifted from 2–12 to 2–15 peptides per protein in the in silico compared to the experimental, respectively (Figure H, ). Importantly, increased peptide per proteins detected with experimental libraries is relevant for peptide-centric analysis, such as proteoform analysis, in which the interpeptide correlations are of high relevance. For example, von Willebrand factor (VWF) is a highly polymorph protein, where polymorphisms can influence its function and dictate clinical phenotype. For this protein, the experimental library (44.3% coverage) allowed for a higher sequence coverage compared to that of the in silico library (33.7%) (Figure I).

Since the in silico library might suffer from the set search space for the prediction of libraries, we then reanalyzed the data from both the in silico and best performing experimental library approach with more restrictive search space. Although the results were identical for the experimental library, we found increased protein identifications along with decreased precursor identifications for the in silico library (). These results highlight the possible benefit of restrictive search space in protein identification in the absence of experimental libraries, although at the risk of lower protein coverage.

In this study, the impact and implications of library creation strategies in neat plasma on protein and peptide identifications were systematically evaluated with a timsTOF HT system. The highest proteomic depth was achieved with the experimental library, closely followed by the in silico library. This high proteomic depth observed for the in silico library is in agreement with previous studies in plasma on a timsTOF HT system. As the in silico libraries indeed generated similar proteomic depth without experimental time, this approach would be a time-efficient approach for protein-centric studies, although the additional computational demands could provide a bottleneck. Interestingly, increased protein coverage was observed when an experimental library was used. In interpretation of these findings it should considered that this study conducted DIA-NN, and as a result the conclusions drawn might be reflective of software performance as different software tools generate their own predicted libraries and use distinct signal extraction and scoring parameters. Concluding, studies that aim to include precursor- or peptide-centric analysis to study proteoforms, for example, will benefit from investing time in building experimental libraries.

Supplementary Material

pr5c00637_si_001.pdf (441.1KB, pdf)
pr5c00637_si_002.xlsx (67KB, xlsx)
pr5c00637_si_003.xlsx (19.5MB, xlsx)

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jproteome.5c00637.

  • Precursor distribution (Figure S1); Precursor distribution faceted by precursor charge (Figure S2); Protein and precursor identifications with adjusted DIA-NN settings (Figure S3); diaPASEF windows of the plasma proteomics method at 30SPD (Table S1); diaPASEF windows of the plasma proteomics method at 60SPD (Table S2); overview of the protein and precursor size across experimental spectral libraries (Table S3); overview of the proteomic depth, data completeness (%), and coefficient of variation (%) in protein quantification for the approaches used to generate spectral libraries (Table S4); Overview of the experimental and computational time dedicated to protein quantification of samples (Table S5) (PDF)

  • Quantified proteins that are shared and unique to experimental or in silico library (Table S6) (XLSX)

  • Proteins in spectral library that are shared and unique to experimental or in silico library (Table S7) (XLSX)

E.R.S., A.J.H., and P.L. designed the experiments, E.R.S., C.v.d.Z., A.J.H., and P.L. performed experiments, E.R.S., A.J.H., and P.L. performed data analysis, E.R.S., S.A.G., and P.L. design figures, E.R.S. and P.L. wrote the manuscript, C.v.d.Z., M.v.d.B., S.A.G., and A.J.H. contributed to the correcting of the manuscript and facilitating the study. P.L. supervised the study. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.

References

  1. Anderson N. L., Anderson N. G.. The Human Plasma Proteome: History, Character, and Diagnostic Prospects. Mol. Cell. Proteomics. 2002;1(11):845–867. doi: 10.1074/mcp.R200007-MCP200. [DOI] [PubMed] [Google Scholar]
  2. Geyer P. E., Holdt L. M.. et al. Revisiting Biomarker Discovery by Plasma Proteomics. Mol. Syst. Biol. 2017;13(9):942. doi: 10.15252/msb.20156297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Deutsch E. W., Omenn G. S.. et al. Advances and Utility of the Human Plasma Proteome. J. Proteome Res. 2021;20(12):5241–5263. doi: 10.1021/acs.jproteome.1c00657. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Lan J., Núñez Galindo A.. et al. Systematic Evaluation of the Use of Human Plasma and Serum for Mass-Spectrometry-Based Shotgun Proteomics. J. Proteome Res. 2018;17(4):1426–1435. doi: 10.1021/acs.jproteome.7b00788. [DOI] [PubMed] [Google Scholar]
  5. Petrera A., von Toerne C.. et al. Multiplatform Approach for Plasma Proteomics: Complementarity of Olink Proximity Extension Assay Technology to Mass Spectrometry-Based Protein Profiling. J. Proteome Res. 2021;20(1):751–762. doi: 10.1021/acs.jproteome.0c00641. [DOI] [PubMed] [Google Scholar]
  6. Rice S. J., Belani C. P.. Characterization of Effective, Simple, and Low-Cost Precipitation Methods for Depleting Abundant Plasma Proteins to Enhance the Depth and Breadth of Plasma Proteomics. Proteomics. 2024;24(15):2400071. doi: 10.1002/pmic.202400071. [DOI] [PubMed] [Google Scholar]
  7. Blume J. E., Manning W. C.. et al. Rapid, Deep and Precise Profiling of the Plasma Proteome with Multi-Nanoparticle Protein Corona. Nat. Commun. 2020;11(1):3662. doi: 10.1038/s41467-020-17033-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Ignjatovic V., Geyer P. E.. Mass Spectrometry-Based Plasma Proteomics: Considerations from Sample Collection to Achieving Translational Data. J. Proteome Res. 2019;18(12):4085. doi: 10.1021/acs.jproteome.9b00503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Skowronek P., Thielert M.. et al. Rapid and In-Depth Coverage of the (Phospho-)­Proteome With Deep Libraries and Optimal Window Design for Dia-PASEF. Mol. Cell. Proteomics. 2022;21(9):100279. doi: 10.1016/j.mcpro.2022.100279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Metatla I., Roger K.. et al. Neat Plasma Proteomics: Getting the Best out of the Worst. Clin. Proteomics. 2024;21(1):22. doi: 10.1186/s12014-024-09477-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Szyrwiel L., Gille C.. et al. Fast Proteomics with Dia-PASEF and Analytical Flow-Rate Chromatography. Proteomics. 2024;24(1–2):e2300100. doi: 10.1002/pmic.202300100. [DOI] [PubMed] [Google Scholar]
  12. Meier F., Brunner A.-D.. et al. DiaPASEF: Parallel Accumulation–Serial Fragmentation Combined with Data-Independent Acquisition. Nat. Methods. 2020;17(12):1229–1236. doi: 10.1038/s41592-020-00998-0. [DOI] [PubMed] [Google Scholar]
  13. Meier F., Beck S.. et al. Parallel Accumulation–Serial Fragmentation (PASEF): Multiplying Sequencing Speed and Sensitivity by Synchronized Scans in a Trapped Ion Mobility Device. J. Proteome Res. 2015;14(12):5378–5387. doi: 10.1021/acs.jproteome.5b00932. [DOI] [PubMed] [Google Scholar]
  14. Meier F., Park M. A.. Trapped Ion Mobility Spectrometry and Parallel Accumulation–Serial Fragmentation in Proteomics. Mol. Cell. Proteomics. 2021;20:100138. doi: 10.1016/j.mcpro.2021.100138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Vitko D., Chou W.-F.. et al. TimsTOF HT Improves Protein Identification and Quantitative Reproducibility for Deep Unbiased Plasma Protein Biomarker Discovery. J. Proteome Res. 2024;23(3):929–938. doi: 10.1021/acs.jproteome.3c00646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Niu L., Thiele M.. et al. Noninvasive Proteomic Biomarkers for Alcohol-Related Liver Disease. Nat. Med. 2022;28(6):1277–1287. doi: 10.1038/s41591-022-01850-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Viode A., Smolen K. K.. et al. Longitudinal Plasma Proteomic Analysis of 1117 Hospitalized Patients with COVID-19 Identifies Features Associated with Severity and Outcomes. Sci. Adv. 2024;10(21):eadl5762. doi: 10.1126/sciadv.adl5762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Zi J., Zhang S.. et al. Expansion of the Ion Library for Mining SWATH-MS Data through Fractionation Proteomics. Anal. Chem. 2014;86(15):7242–7246. doi: 10.1021/ac501828a. [DOI] [PubMed] [Google Scholar]
  19. Gillet L. C., Navarro P., Tate S., Rost H., Selevsek N., Reiter L., Bonner R., Aebersold R.. et al. Targeted Data Extraction of the MS/MS Spectra Generated by Data-Independent Acquisition: A New Concept for Consistent and Accurate Proteome Analysis. Mol. Cell. Proteomics. 2012;11(6):O111.016717. doi: 10.1074/mcp.O111.016717. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Govaert E., Van Steendam K., Willems S., Vossaert L., Dhaenens M., Deforce D.. Comparison of Fractionation Proteomics for Local SWATH Library Building. Proteomics. 2017;17(15–16):1700052. doi: 10.1002/pmic.201700052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Demichev V., Messner C. B.. et al. DIA-NN: Neural Networks and Interference Correction Enable Deep Proteome Coverage in High Throughput. Nat. Methods. 2020;17(1):41–44. doi: 10.1038/s41592-019-0638-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Zeng W.-F., Zhou X.-X.. et al. AlphaPeptDeep: A Modular Deep Learning Framework to Predict Peptide Properties for Proteomics. Nat. Commun. 2022;13(1):7238. doi: 10.1038/s41467-022-34904-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Nigjeh E. N., Chen R.. et al. Quantitative Proteomics Based on Optimized Data-Independent Acquisition in Plasma Analysis. J. Proteome Res. 2017;16(2):665–676. doi: 10.1021/acs.jproteome.6b00727. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li J., Jiang X.-J.. et al. Data-Independent Acquisition Proteomics Reveals Circulating Biomarkers of Coronary Chronic Total Occlusion in Humans. Front. Cardiovasc. Med. 2022;9:960105. doi: 10.3389/fcvm.2022.960105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Jin L., Wang F.. Identification of Plasma Biomarkers from Rheumatoid Arthritis Patients Using an Optimized Sequential Window Acquisition of All THeoretical Mass Spectra (SWATH) Proteomics Workflow. Proteomes. 2023;11(4):32. doi: 10.3390/proteomes11040032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Messner C. B., Demichev V., Wendisch D., Michalick L., White M., Freiwald A., Textoris-Taube K., Vernardis S. I., Egger A.-S., Kreidl M.. et al. Ultra-High-Throughput Clinical Proteomics Reveals Classifiers of COVID-19 Infection. Cell Syst. 2020;11(1):11–24.e4. doi: 10.1016/j.cels.2020.05.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kverneland A. H., Østergaard O.. et al. Differential Ultracentrifugation Enables Deep Plasma Proteomics through Enrichment of Extracellular Vesicles. Proteomics. 2023;23(7–8):2200039. doi: 10.1002/pmic.202200039. [DOI] [PubMed] [Google Scholar]
  28. Kulak N. A., Pichler G.. et al. Minimal, Encapsulated Proteomic-Sample Processing Applied to Copy-Number Estimation in Eukaryotic Cells. Nat. Methods. 2014;11(3):319–324. doi: 10.1038/nmeth.2834. [DOI] [PubMed] [Google Scholar]
  29. Teo G. C., Polasky D. A.. et al. Fast Deisotoping Algorithm and Its Implementation in the MSFragger Search Engine. J. Proteome Res. 2021;20(1):498–505. doi: 10.1021/acs.jproteome.0c00544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Yu F., Haynes S. E.. et al. IonQuant Enables Accurate and Sensitive Label-Free Quantification With FDR-Controlled Match-Between-Runs. Mol. Cell. Proteomics. 2021;20:100077. doi: 10.1016/j.mcpro.2021.100077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kong A. T., Leprevost F. V.. et al. MSFragger: Ultrafast and Comprehensive Peptide Identification in Mass Spectrometry–Based Proteomics. Nat. Methods. 2017;14(5):513–520. doi: 10.1038/nmeth.4256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Yu F., Haynes S. E.. et al. Fast Quantitative Analysis of TimsTOF PASEF Data with MSFragger and IonQuant. Mol. Cell. Proteomics. 2020;19(9):1575–1585. doi: 10.1074/mcp.TIR120.002048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Demichev V., Szyrwiel L.. et al. Dia-PASEF Data Analysis Using FragPipe and DIA-NN for Deep Proteomics of Low Sample Amounts. Nat. Commun. 2022;13(1):3944. doi: 10.1038/s41467-022-31492-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Bludau I., Frank M.. et al. Systematic Detection of Functional Proteoform Groups from Bottom-up Proteomic Datasets. Nat. Commun. 2021;12(1):3810. doi: 10.1038/s41467-021-24030-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kreft I. C., van Duijl T. T.. et al. Variant Mapping Using Mass Spectrometry-Based Proteotyping as a Diagnostic Tool in von Willebrand Disease. J. Thromb. Haemost. 2024;22(7):1894–1908. doi: 10.1016/j.jtha.2024.04.011. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

pr5c00637_si_001.pdf (441.1KB, pdf)
pr5c00637_si_002.xlsx (67KB, xlsx)
pr5c00637_si_003.xlsx (19.5MB, xlsx)

Data Availability Statement

The mass spectrometry raw data (.d files) and processed data (.tsv) have been deposited in the ProteomeXchange/PRIDE archive database with identifier PXD058337.

Articles from Journal of Proteome Research are provided here courtesy of American Chemical Society

RESOURCES