henhaobio
Sign in

Nature BiotechnologyBioengineeringPeer-reviewed journal article

Increasing the throughput of sensitive proteomics by plexDIA

Jason DerksiD, Andrew LeduciD, Georg WallmanniD, R. Gray HuffmaniD, Matthew Willetts, Saad KhaniD, Harrison SpechtiD, Markus RalseriD, Vadim DemicheviD, Nikolai SlavoviD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

Current mass-spectrometry methods enable high-throughput proteomics of large sample amounts, but proteomics of low sample amounts remains limited in depth and throughput. To increase the throughput of sensitive proteomics, we developed an experimental and computational framework, plexDIA, for simultaneously multiplexing the analysis of both peptides and samples. Multiplexed analysis with plexDIA increases throughput multiplicatively with the number of labels without reducing proteome coverage or quantitative accuracy. By using 3-plex nonisobaric mass tags, plexDIA enables quantifying 3-fold more protein ratios among nanogram-level samples. Using 1 hour active gradients and first-generation Q Exactive, plexDIA quantified about 8,000 proteins in each sample of labeled 3-plex sets. plexDIA also increases data completeness, reducing missing data over 2-fold across samples. We applied plexDIA to quantify proteome dynamics during the cell division cycle in cells isolated based on their DNA content; plexDIA detected many classical cell cycle proteins and discovered new ones. When applied to single human cells, plexDIA quantified about 1,000 proteins per cell and achieved 98 % data completeness within a plexDIA set while using about 5 min of active chromatography per cell. These results establish a general framework for increasing the throughput of sensitive and quantitative protein analysis. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=81 SRC="FIGDIR/small/467007v2_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@144bba6org.highwire.dtl.DTLVardef@17bbf20org.highwire.dtl.DTLVardef@140f497org.highwire.dtl.DTLVardef@179faa9_HPS_FORMAT_FIGEXP M_FIG C_FIG

In-depth summary

Claude Opus 5.5 · xhigh effort · 6 Nov 2026 · from the full text of version 2 (22 Mar 2022) · CC BY-NC-ND 4.0

Multiplexing nonisobaric isotopologous mass tags (mTRAQ) with data-independent acquisition and DIA-NN increases sensitive proteomics throughput 3-fold while cutting missing data below 3% in bulk and single cells.

Proteomics of nanogram-level and single-cell samples is bottlenecked by long ion accumulation times that limit serial data-dependent acquisition (DDA) throughput, while label-free data-independent acquisition (LF-DIA) analyzes only one sample per run. Combining parallel precursor isolation in DIA with 3-plex nonisobaric mTRAQ chemical labeling and retention-time-aligned channel propagation in DIA-NN multiplies sample throughput without sacrificing proteome depth or quantitative accuracy.

Key results

  1. Shown3-plex plexDIA multiplicatively increases proteome quantification throughput and reduces missing data compared to label-free DIA on 500 ng mixed-species standards.

    Using 60-min active gradients on a first-generation Q-Exactive, 3-plex plexDIA quantified 6,282 (V1) and 7,923 (V2) intersecting protein groups across 3 samples versus 5,851 (V1) and 8,318 (V2) for LF-DIA, increasing protein ratio quantification rates by 3.22-fold (V1) and 2.86-fold (V2) per unit time and reducing within-set missing data to 2-3% vs 13-15% across replicate injections. · plexDIA benchmarks; plexDIA increases throughput multiplicatively; plexDIA increases data completeness across…; Figure legends

  2. ShownQuantitative accuracy, repeatability, and differential abundance detection of plexDIA match label-free DIA at 3-fold lower instrument time.

    Across 12,863 sample-specific protein groups quantified in triplicate, median CVs were 0.103 for plexDIA and 0.108 for LF-DIA; U-937 vs Jurkat protein ratios showed Spearman r=0.90 across 1,078 differentially abundant proteins (1% FDR) and r=0.78 across all 2,728 proteins, while recalling 643 vs 663 true-positive yeast/bacterial proteins at 1% empirical FDR. · Quantitative accuracy of plexDIA is comparable…; Repeatability of plexDIA is comparable to LF-DIA; Estimating differential protein abundance by…; Figure legends

  3. ShownSingle-cell plexDIA quantifies ~1,000 proteins per human cell with 98% within-set data completeness using 5 minutes of active chromatography per cell.

    Across 3-plex single-cell sets of U-937, PDAC (HPAF-II), and melanoma (WM989-A6-G3) cells prepared via nPOP and analyzed on timsTOF SCP (30 cells, 10 sets) and Q-Exactive (132 cells, 12 negative controls, 48 sets; 155 cells passing <60% missingness QC), plexDIA achieved >98% within-set completeness on timsTOF SCP and >50% between-set completeness on Q-Exactive across a 1,000-fold dynamic range. · Abstract; Single-cell analysis with plexDIA; Preparation of single-cell plexDIA samples; Figure legends

  4. ShownApplying plexDIA to FACS-sorted U-937 cells resolves canonical and previously uncharacterized cell division cycle protein dynamics without chemical synchronization.

    Analyzing ~2,000 cells per phase (G1, S, and G2/M sorted by DNA content) with 2-hour V1 and V2 gradients quantified 4,391 unique protein groups (4,107 gene groups) and 400 differentially abundant proteins at 1% FDR, identifying MCM complex, polo-like kinase 1, ubiquitin-conjugating enzyme E2, TCA cycle and fatty acid beta-oxidation pathways, plus CDV3 and JPT2. · Cell division cycle analysis with plexDIA; Cell division cycle, FACS and sample preparation; Figure legends

Methods, data and reproducibility

Samples are labeled with 3-plex nonisobaric mTRAQ tags (Δ0, Δ4, Δ8), pooled, and acquired via MS1-optimized (V1), MS2-optimized (V2), or dia-PASEF DIA on Q-Exactive or timsTOF SCP instruments. A new plexDIA module in DIA-NN uses neural networks and decoy channels to propagate peptide IDs across co-eluting isotopologues, quantifying via apex fragment ratios and correcting 4 Da arginine isotopic interference.

  • Mixed-species bulk proteome standards (Samples A, B, and C mixing E. coli, S. cerevisiae, and H. sapiens U-937 and Jurkat lysates, 500 ng per sample) · multiple species · n = 3 samples (A, B, C) run in triplicate across V1 and V2 methods · MSV000089093
  • Two-species spiked benchmark standard for decoy-channel FDR validation (0.5 µg E. coli + 0.5 µg S. cerevisiae in mTRAQ Δ0 vs 0.5 µg S. cerevisiae in mTRAQ Δ4) · multiple species · n = 2 channels in 1 plexDIA set · MSV000089093
  • FACS-sorted U-937 monocyte cell division cycle fractions (G1, S, and G2/M phases isolated by Vybrant DyeCycle Violet DNA staining) · human · n = 3 fractions (~2,000 cells per cell-cycle phase, 1 plexDIA set) · MSV000089093
  • Single human cells from U-937 monocyte, HPAF-II PDAC, and WM989-A6-G3 melanoma lines prepared by nPOP droplet workflow, plus 100-cell bulk standards · human · n = 132 cells on Q-Exactive (48 sets) and 30 on timsTOF SCP (10 sets); 155 passed QC · MSV000089093
Methods & tools
DIA-NN (plexDIA module), MaxQuant, MaxLFQ, mTRAQ, nPOP, mPOP, ThermoRawFileParser, CellenONE, timsTOF SCP (dia-PASEF), Q-Exactive
Software & environment
DIA-NN v1.8.1 beta 16, MaxQuant v1.6.17.0, DIA-NN R package, modified ThermoRawFileParser, Swiss-Prot H. sapiens, E. coli, and S. cerevisiae FASTAs (February 2022)
Limitations (5)
  • Commercially available nonisobaric tags for plexDIA are currently limited to low-plex capacity (such as 3-plex mTRAQ), compared with 16-plex or 18-plex isobaric TMTpro tags used in DDA. · stated by the authors
  • C-terminal arginine peptides carry a single mTRAQ label separated by 4 Da, yielding only sample-specific b-ions at MS2, slightly lower MS2 quantification accuracy than lysine peptides, and requiring theoretical isotopic envelope subtraction at MS1. · stated by the authors
  • Scaling plexDIA beyond 3 labels increases spectral complexity and potential ion interference and is constrained by finite MS detector ion capacity unless narrower m/z windows or higher resolving power are used. · stated by the authors
  • The MS2-optimized duty cycle (V2) achieved deeper proteome coverage than the MS1-optimized cycle (V1) but resulted in lower quantitative accuracy for both LF-DIA and plexDIA, and one LF-DIA V2 replicate of sample C was excluded as an outlier. · stated by the authors
  • Single-cell sets acquired on the timsTOF SCP and Q-Exactive were prepared on different days from different cell batches, and about 5% of single cells failed sample preparation (>60% missing precursor data). · stated by the authors

Read this if you develop or apply mass-spectrometry workflows for low-input or single-cell proteomics and want to multiply sample throughput and data completeness using multiplexed DIA.

Citations
235
FWCI
17.51
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISDerks '23Derks '25Specht '25Thielert '23Guzmán '23Richards '22
87% semantic matchBioengineering
Jason Derks, Nikolai Slavov·Journal of Proteome Research·Feb 3, 2023· 29 citations