BGPT: Paper Review: DIA-MS based plasma peptidomic workflow for profiling organ-derived peptides

Explore by Goal

Quick Explanation Copied

DIA-PO (empirical organ-derived spectral library) outperforms DDA for mouse plasma peptidomics: ~5,587 vs ~4,394 peptides identified, and organ-derived peptides ~2,108 vs ~1,287 (all reported in the paper’s Results) .

Long Explanation

DIA-MS based plasma peptidomic workflow for profiling organ-derived peptides — rigorous visual critique

Mouse organ→plasma mapping via DIA-NN using an empirical organ-derived spectral library (DIA-PO). Evidence and numerical claims are limited to what is explicitly stated in the provided paper text and reference list.

What the paper claims (constrained to text)

Empirical spectral library strategy: peptides extracted from plasma and 11 organs; DDA-MS is used to build libraries; DIA-NN is used to analyze plasma DIA .
Quantitative performance: DIA-PO identifies ~5,587 peptides vs DDA ~4,394 (≈1.3×) from an equivalent of ~6 µL mouse plasma (triplicate) .
Organ mapping: DIA-PO maps 2,683 organ-derived peptides to their organs of origin across 11 organs; many appear to be observed in only a single organ .
Bioactive peptide examples: 19 known bioactive peptides (annotated as “Peptide” in UniProt) are included; neuropeptide Y (NPY) is reported as full-length identified in plasma by non-targeted analysis .

VISUAL 1 — Identification yields (DIA-PO vs DDA)

Source numbers are directly from the paper’s Results text .

VISUAL 2 — Replicate quantification coverage

Replicate-count metrics are explicitly reported by the paper .

VISUAL 3 — Library designs compared (library-source logic)

The figure is a schematic derived from the paper’s workflow definitions: empirical libraries (“Plasma”, “Plasma and Organs”) and “InfinDIA”, then DIA-NN analysis of plasma .

Critical interpretation (what the numbers do and do not prove)

1) Library construction plausibly affects sensitivity, but “organ-derived” mapping depends on assumptions.

Known from the text: DIA-PO uses an empirical library built from plasma plus 11 organ DDA-MS datasets (FragPipe/MSFragger NONSPECIFIC peptidome workflow with peptide lengths 7–50) .
Uncertain/limitation: The organ library is built from 11 organs measured once (explicitly stated), without reducing/alkylating cysteine thiol groups, which can bias the proteoform spectrum captured in the library .

2) Improved counts can reflect better detectability of low-intensity peptides, not necessarily higher biological specificity.

The paper reports DIA-PO-unique peptides have lower signal intensities than peptides detected by both DIA-PO and DDA, and that this is more pronounced for those also observed in organs—consistent with improved sensitivity for low abundance species .
But: Increased library matching can reduce false negatives; it does not automatically validate organ origin causality. Mapping remains model- and library-dependent.

3) Transport/processing in blood is a major confound for “organ-of-origin”.

The paper observes frequent truncated peptides lacking amino acids at either N- or C-terminus, and argues this suggests further processing/degradation in circulation .
Why this matters: A peptide’s presence in plasma may represent a mixture of organ secretion, plasma protease processing, and clearance, so organ mapping is at best a probabilistic inference unless validated orthogonally.

Methodological audit (from provided Methods text)

Acquisition details:

They use timsTOF HT with ddaPASEF for empirical spectral library generation, and diaPASEF (modified Thin-diaPASEF) for DIA analysis .
For diaPASEF: 31 windows, isolation width 25 Th, 1 Th overlap, ramp/accumulation 75 ms (cycle ~1.0 s) and denoising set to “Low Sample Amount (Sensitive)” .

Search/library analysis:

Library generation uses FragPipe v23.0 with nonspecific peptidome workflow; peptide length 7–50; NONSPECIFIC cleavage; variable modifications include oxidation(M), acetylation(N-term), amidation(C-term), and pyro-Glu from Q/E; validation via MSBooster/Percolator and peptide FDR 1% .
DIA analysis uses DIA-NN v2.3.1 with InfinDIA pre-search, “Prediction from FASTA”, N-term M excision, acetyl(N-term), oxidation(M), pyro-glu options; precursor FDR 1% .

Data availability + reproducibility posture

They deposit MS data: ProteomeXchange PXD073758 and jPOST JPST003835 .
Reproducibility caveat from the paper itself: organ-derived library coverage is limited to 11 organs, each measured once; future expansion with replicate organ libraries and cysteine reduction/alkylation is expected to improve coverage .

Author reviews (bespoke links)

Feedback:

Updated: April 12, 2026