Decision letter

Decision Letter

VERDICT: major

Summary of Evaluation

This manuscript presents RIPUP, a rapid multi-protease workflow (Arg-C Ultra plus r-Chymotrypsin, with optional TMT derivatization) for bottom-up histone PTM analysis, benchmarked against the field-standard Trypsin/propionylation protocol and analysed through the HiP-Frag search framework. The systematic design — three proteases, ten conditions, forty samples, consistent metrics for CV, digestion efficiency, sequence coverage, labeling efficiency and PTM diversity — is genuinely useful, and the field has needed exactly this kind of controlled head-to-head comparison. Data and code are deposited with specific accessions (PXD073683; GitHub), search parameters are largely specified, ethics and funding statements are complete, and the authors are candid about several limitations. Two findings in particular are well supported and independently valuable: Arg-C Ultra's high cleavage specificity and low missed-cleavage rate, and the demonstration that propionylation efficiency in the widely used ammonium-bicarbonate protocol is far lower than commonly assumed (29–71%), with the downstream consequence that a large fraction of tryptic peptides is invisible to the standard search specificity.

All five specialist reports converged on a score of 4, and the panel is in agreement that the work is sound and worth publishing. Where I part company with a straight "minor" reading is on two specific points, both of which are load-bearing for claims that appear in the Abstract and Conclusions, and both of which require analysis whose outcome could change what the paper concludes.

First, the flagship claim — that TMT's tertiary amine "provides charge compensation that rescues the ionization of negatively charged acylations", revealing a "dark epigenome" of 58 succinylation and 31 glutarylation sites — is confounded by the acquisition method stated in your own Methods: non-TMT samples were fragmented at a fixed 30% NCE, TMT samples at stepped 30/40/50% NCE. Three reviewers arrived at this independently from different entry points, and the advocate in the panel debate conceded the confound is real and unresolved for isolating the mechanism. More thorough fragmentation, higher labeling efficiency (99% vs ~68% by intensity), and altered chromatographic behaviour are all live alternative explanations for the differential site counts. The observation that TMT-labeled Arg-C Ultra digests yield many more acidic acylation identifications is intact and interesting; the causal mechanism as currently asserted is not established by this design. Importantly, a partial discriminating analysis is already available in your data: unlabeled Arg-C Ultra and r-Chymotrypsin digests were acquired at fixed NCE and retain free N-terminal amines. Their succinylation/glutarylation counts (not currently reported separately) would help separate "charge at the N-terminus" from "TMT structure specifically" from "stepped energy". That analysis could reasonably change the framing of the paper's headline result, which is why this is a major rather than a minor revision.

Second, the NAM quantitative experiment rests on peptidoform-level intensities in a setting where you yourselves document dose-dependent missed-cleavage redistribution (259 peptidoforms detected only in treated samples). You correctly rejected peptide-family ratios for this reason, but the chosen alternative does not remove the confound: a peptidoform can rise with dose because the site is more modified or because the backbone became more detectable. You report this explicitly for r-Chymotrypsin (12 of 16 significant peptidoforms unmodified) but not for the 112 significant Arg-C Ultra peptidoforms that support the claim of quantitative performance. Stratifying those 112 and repeating the contrasts on zero-missed-cleavage peptidoforms only is a bounded reanalysis, but its outcome bears directly on whether Figure 7 demonstrates PTM quantitation or partly cleavage-state redistribution.

Beyond these, there is a plain internal contradiction between the Abstract ("exceeds Trypsin-based approaches") and the Results ("comparable PTM numbers to conventional 'Trypsin + Prop' methods (~120 PTMs)"), and the protease comparison is systematically confounded with labeling chemistry — the one same-label contrast available (Arg-C Ultra + Prop, 254 peptides vs Trypsin + Prop, 179) shows a considerably more modest advantage than the framing implies. The compliance audits also flag several HARD reporting gaps (cell line provenance and authentication, randomization/blinding statement for the rat work, verification that the delegated histone-extraction protocol is self-contained, fragment mass tolerance, unenumerated deviations from the published HiP-Frag workflow). None of these gaps is fatal, and none prevents evaluation of the central claims, but all must be closed.

I want to be clear about what this verdict is not. It is not a judgement that the workflow is unsound, that the succinylation/glutarylation observations are spurious, or that synthetic-peptide validation of every new site is required before publication. Your own position — following Vai et al. in using rigorous computational filtering with targeted validation reserved for ambiguous cases — is a defensible one for a methods paper, and I am not requiring orthogonal validation of the 89 acidic acylation sites. What I am requiring is that the causal mechanistic claim be either supported by matched-energy evidence or restated as the hypothesis it currently is, and that the quantitative claim be shown to survive the cleavage-redistribution confound you identified.

Required Revisions

1. Resolve or reframe the collision-energy confound underlying the charge-compensation claim. Either (a) acquire matched-energy data — propionylated Arg-C Ultra at stepped 30/40/50% NCE and/or TMT-labeled Arg-C Ultra at fixed 30% NCE — and report succinylation/glutarylation site counts for the matched comparison; or (b) if re-acquisition is not feasible, state the confound explicitly in the Results where the finding is introduced (not only in Limitations), and revise the Abstract, the Figure 3 narrative and the Conclusions so that charge compensation is presented as a mechanistic hypothesis consistent with the data rather than as a demonstrated cause. Phrases such as "TMT's tertiary amine provides charge compensation that rescues the ionization" must be requalified under option (b).

2. Report the already-available discriminating comparison. From the existing dataset (Figure 5A/5B), tabulate succinylation and glutarylation site counts for unlabeled Arg-C Ultra, unlabeled r-Chymotrypsin, propionylated Arg-C Ultra and r-Chymotrypsin, and TMT-labeled digests, all with the acquisition NCE stated per condition. Discuss what the unlabeled (fixed-NCE, free N-terminal amine) counts imply for the relative contributions of N-terminal charge, TMT structure and fragmentation energy. Also report whether the reported acidic acylation sites overlap between the two proteases.

3. Reanalyse the NAM quantitative experiment against the missed-cleavage confound. (i) Stratify the 112 FDR-significant Arg-C Ultra peptidoforms into modified and unmodified, and report the counts, as you already do for r-Chymotrypsin; (ii) repeat the limma contrasts restricted to peptidoforms with zero missed cleavages and report whether the sirtuin-target directionality (H3 K9ac/K23ac, H4 K12ac/K16ac, H4 K44ac) survives; (iii) state which conclusions in this section depend on the unrestricted analysis. If the restricted analysis weakens the result, say so and scale the claim accordingly.

4. Reconcile the Abstract with the Results on total PTM detection. Provide the exact unique-PTM-site count per condition (a table or explicit numbers in the Figure 5 caption), and rewrite the Abstract sentence so it matches the data — e.g. comparable total PTM detection to Trypsin + Prop with unique access to acidic acylations, if that is what Figure 5A shows. The same correction applies to the Conclusions.

5. Separate protease effect from labeling effect in the benchmarking claims. Foreground Arg-C Ultra + Prop vs Trypsin + Prop (254 vs 179 fully cleaved peptides) as the controlled protease comparison, state plainly that Arg-C Ultra + TMT vs Trypsin + Prop conflates protease and labeling chemistry, and temper "superior" / "exceeds" language for the protease itself to what the same-label contrast supports. Your existing justification for not running Trypsin + TMT is reasonable and should be retained, but it does not license the stronger framing.

6. Quantify the dual-protease complementarity claim. For both HEK293T and rat hippocampus, report unique-to-Arg-C Ultra, unique-to-r-Chymotrypsin and shared PTM sites. For the rat data (n = 5), report the detection-frequency distribution across animals for the 231 sites (how many in 5/5, 4/5, 3/5, 2/5), and restate the claim as "231 sites detected in ≥2 of 5 animals" with that breakdown visible.

7. Complete the quantitative-analysis transparency reporting. State: the number and location of imputed values; the fraction of the 112 (and 16) significant peptidoforms containing any imputed value; a complete-case sensitivity analysis; the CV threshold used for filtering and how many peptidoforms it removed; and the random seed (or seed-averaging) for kNN imputation. Report specific versions of limma and impute.

8. Complete the search-space and FDR reporting. Report the number of PSMs and peptides passing 1% FDR per condition; summarise the size of the variable-modification and mass-offset space used (SI Table S1 should be summarised in the main text, not only referenced); state the fragment mass tolerance used in the FragPipe searches; and enumerate the specific deviations covered by "following the recommended guidelines for the HiP-Frag workflow, with some modifications".

9. Close the HARD reagent and design-reporting gaps flagged by the methods audit. Add HEK293T provenance with RRID/CVCL, an authentication statement (or an explicit statement that STR profiling was not performed), and a mycoplasma-testing statement; add a randomization/blinding statement for the rat tissue collection and analysis, or state explicitly that neither was applied and why; and confirm in the main text that SI Methods contains a complete, self-contained histone-extraction protocol (if it delegates further, provide the full protocol).

10. Temper the formylation observation. You cite ref 60 showing formylation can arise from formic-acid exposure and then assert "this was not a feature of our workflow" without supporting evidence. Either provide the basis for that assertion (e.g. a formic-acid-free control, or the exposure conditions that make the artifact implausible here) or present formylation as an observation of uncertain origin.

11. State the propionylation-benchmark caveat where the comparison is made. The acknowledgement that ammonium-containing buffers depress propionylation efficiency currently appears only in the Figure 4 caption and Discussion. Add it to the Results where Figure 4A is first interpreted, and state clearly that the TMT-vs-propionylation efficiency gap is measured against the widely used protocol rather than an optimised (e.g. TEAB-based) propionylation.

12. Fix reference-list indexing. Entries 43–59 appear as an unnumbered block; renumber so each in-text citation maps unambiguously to one reference.

Minor Suggestions

• Add caveats to the cost analysis (SI Table S5): instrument time is excluded, and r-Chymotrypsin pricing is not established under Early Access. Note also that reagent availability may limit near-term reproduction of the dual-protease arm.
• Distinguish your contribution from Ryzhaya et al. (ref 10) more explicitly in the Introduction or Discussion — one or two sentences naming what RIPUP adds beyond Arg-C Ultra plus peptide-level TMA derivatization (TMT chemistry, dual-protease coverage, tissue application, unrestricted search) would help readers place the work.
• Report statistical support (effect size or enrichment p-value) for the missed-cleavage motif enrichments in SI Figures S6–S7; as presented it is unclear whether D at P1′ and E at P2 are strong or marginal associations.
• For Figure 4A, note how many peptides contribute to the intensity-weighted efficiency and whether a small number of abundant peptides dominate — this bears on whether ~99% efficiency translates to rare, low-abundance PTM-bearing peptides.
• Quantify the endogenous propionylation/butyrylation detected in unlabeled digests, or state the extent to which they are expected to affect the unlabeled Arg-C Ultra and r-Chymotrypsin results.
• The suggestion that DIA could be layered onto the HiP-Frag framework is worth keeping, but consider stating what sensitivity gain you would expect and why, so it reads as a concrete extension rather than an aside.
• Validating the acidic-acylation enhancement in a second cell type or tissue, and targeted PRM or synthetic-peptide confirmation of the most abundant succinylation/glutarylation sites, would substantially strengthen the "dark epigenome" framing in future work. I am not requiring either here, but flagging them as the obvious next steps if the claim is to be generalised.