Pith. sign in

REVIEW 2 major objections 4 minor 6 references

Train-by-train scale-plus-offset fits plus matched water controls recover a reproducible, transmission-dependent residual SAXS signal in aqueous cysteine that averages near 1/√N.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-14 12:24 UTC pith:XFSBFBBX

load-bearing objection Solid methods paper: train-level scale-plus-offset plus water–water controls recovers a reproducible, fluence-dependent residual SAXS signal with near-1/√N statistics; origin left open and isolation from residual run mismatch is the softest point. the 2 major comments →

arxiv 2607.10349 v1 pith:XFSBFBBX submitted 2026-07-11 physics.optics

Train-Resolved Statistical Recovery of Weak SAXS Signals in Liquids at the European XFEL

Carles Serrat , Asier Garc\'ia , Biel Serrat , Angelo Beratto-Ramos , Johan Bielecki , Huijong Han , Sara Hern\'andez , Tokushi Sato
show 6 more authors
Joana Valerio Mohammad Vakili Egor Sobolev Katerina Doerner Chan Kim Majed Chergui
This is my paper
classification physics.optics
keywords SAXSXFELtrain-resolved analysisL-cysteineresidual scatteringscale-plus-offsetwater controlshigh-repetition-rate
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

High-repetition-rate XFELs promise sensitivity to weak scattering through massive averaging, but detector-wide fluctuations, jet variability, and normalization drift often prevent ideal 1/√N behaviour. This paper shows that reconstructing ON–OFF radial profiles per XFEL train, fitting each cysteine–water train pair with its own scale and offset, and then subtracting transmission-matched water–water controls isolates a residual that conventional averaged difference analysis buries. In 0.5 M L-cysteine the residual is sign-changing (positive near 0.05–0.15 Å⁻¹, negative near 0.16–0.30 Å⁻¹), grows with incident transmission, and emerges progressively as independent train pairs are accumulated. Block-to-block variability of the integrated residual tracks the expected inverse-square-root scaling with block size. A sympathetic reader cares because the same protocol converts the European XFEL’s train structure into genuine statistical leverage for any weak solution-scattering signal once common-mode contributions are removed.

Core claim

After independent train-by-train scale-plus-offset subtraction of cysteine versus water and removal of transmission-matched water–water control residuals, the 0.5 M cysteine dataset retains a reproducible sign-changing residual SAXS structure that strengthens with incident XFEL transmission and whose integrated observables show block-to-block variability consistent with approximately 1/√N averaging over independent train pairs.

What carries the argument

The control-corrected residual ΔR(q) = R_cys−water(q) − R_water−water(q), built after independent scale-plus-offset fitting I_cys(q) = a I_water(q) + b for every matched train pair. It removes detector-wide multiplicative and additive contributions and subtracts the residual structure measured when identical water samples are compared under the same conditions, leaving only structure that exceeds those reproducibility limits.

Load-bearing premise

That transmission-matched water–water residuals fully capture all non-sample common modes, so the cysteine–water minus water–water difference isolates a sample-related contribution rather than leftover run-to-run mismatch.

What would settle it

A fully alternated water–cysteine–water sequence at the same three transmissions that yields a flat control-subtracted residual, or integrated A+ and A− values that fall inside the water-control uncertainty, would falsify the claim that a sample-dependent residual survives the analysis.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • High-repetition-rate XFEL operation can be converted into real statistical sensitivity for weak SAXS once train-level ON–OFF observables and matched controls suppress common-mode fluctuations.
  • Weak residual signals hidden in conventional averaged comparisons can emerge after train-level reconstruction and accumulation of large train-pair ensembles.
  • Future experiments with fully alternated water/sample sequences, denser transmission sampling, and improved control statistics can determine the physical origin of the residual.
  • Similar statistical considerations apply to other XFEL measurements that target intrinsically weak signals, including nonlinear X-ray processes.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same per-train scale-plus-offset plus matched-control pipeline should generalize to other dilute biomolecules or solvents where fluence-dependent structural or radiolytic effects are suspected but previously undetectable.
  • If the residual is beam-induced, denser fluence sampling could map a threshold or power-law response and help separate density fluctuations from chemical product formation.
  • Alternating sample–reference acquisition should become standard practice for any XFEL-SAXS campaign that aims at sub-percent residuals.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The manuscript introduces a train-resolved SAXS analysis pipeline for high-repetition-rate XFEL data and applies it to aqueous 0.5 M L-cysteine solutions measured at the European XFEL SPB/SFX instrument with AGIPD. For each XFEL train an energy-normalized ON–OFF radial profile is formed; matched cysteine–water train pairs are then fit independently by a scale-plus-offset model Icys(q)=a Iwater(q)+b, after which the residual R(q) is corrected by subtracting the analogous residual obtained from transmission-matched water–water controls. The resulting control-subtracted residual ΔR(q) exhibits a reproducible sign-changing shape (positive lobe ~0.05–0.15 Å⁻¹, negative lobe ~0.16–0.30 Å⁻¹) that is weak at T=0.0099 and strengthens at T=0.21 and 0.32. Integrated observables A+ and A−, convergence tests over increasing numbers of train pairs, and block-averaging statistics are used to show that the residual is an ensemble property whose block-to-block variability scales approximately as 1/√N. The microscopic origin is left unresolved; the principal claims are the existence of a statistically robust, transmission-dependent residual that survives the stated corrections and the demonstration that train-level observables plus matched controls can recover near-ideal statistical averaging for weak SAXS signals.

Significance. If the residual is real and the analysis pipeline is robust, the work supplies a concrete, transferable protocol for extracting weak fluence-dependent SAXS contrasts from megahertz XFEL liquid-jet data—precisely the regime in which conventional averaged difference profiles are dominated by common-mode fluctuations. The explicit recovery of ~1/√N scaling after train-resolved scale-plus-offset correction and water–water subtraction is a useful methodological result for the XFEL community. Strengths that should be credited include the conservative scope (origin left open), the use of transmission-matched water–water controls, the progressive-convergence and block-averaging tests (Figs. 3–4), and the decision to exclude imperfectly controlled data sets (T=0.046 and 1 M). These elements make the paper a solid contribution to experimental methodology even if the physical mechanism remains unidentified.

major comments (2)
  1. [Sections III–IV, Table I] The central isolation claim—that ΔR(q) exceeds non-sample reproducibility limits—rests on the premise that transmission-matched water–water residuals fully capture the common-mode contributions present in the cysteine–water pairs (Sections III–IV). Table I, however, shows that cysteine and water runs form sequential blocks rather than interleaved acquisitions (e.g., Cys 24–25 vs Water 22–23 at T=0.21; Cys 18–19 vs Water 20–21 at T=0.32). Residual slow drifts in jet diameter/velocity, beam pointing or detector response between those blocks can therefore leave a non-sample contribution in ΔR that is larger than the pure water–water residual. The manuscript already excludes T=0.046 and the 1 M data for precisely this reason; the same logic implies that imperfect temporal matching is the softest point of the retained data sets. A quantitative bound on inter-block variability (for example by
  2. [Section III] The scale-plus-offset parameters a and b are obtained by least-squares minimization “over the selected fitting range” (Section III), yet that q-range is never stated. Because the residual lobes that define A+ and A− lie inside the SAXS window, the numerical values of a and b—and therefore the amplitude and even the sign pattern of R(q)—can depend on whether the fit includes or excludes those lobes. The same ambiguity applies to the water–water controls. The manuscript should specify the exact fitting interval, demonstrate that the reported ΔR(q) shape and the transmission trend of A+/A− are stable under reasonable variations of that interval, and confirm that the identical interval was used for all cysteine–water and water–water pairs.
minor comments (4)
  1. [Section V.C, Table II] The integration windows that define A+ (0.05–0.15 Å⁻¹) and A− (0.16–0.30 Å⁻¹) are chosen after inspection of the residual shape (Figs. 1–2). A short sensitivity test (shifting the windows by ±0.02 Å⁻¹ or using the full 0.05–0.30 Å⁻¹ interval) would show that the transmission trend is not an artifact of post-hoc window selection.
  2. [Figures 1–3 captions] Figures 1–3 note that curves are “lightly smoothed for visual clarity” while quantitative analysis uses unsmoothed data. Please state the smoothing kernel (or omit smoothing entirely) so that readers can judge whether any visual features are introduced by the display step.
  3. [Section V, Table II] Units of the residual intensity and of the integrated observables A+/A− are never given. Even if the profiles are arbitrarily scaled after the a,b fit, a brief statement (e.g., “relative to the mean water intensity in the fit window”) would aid comparison with future measurements.
  4. [Section V.E] The Pearson correlation ρ=0.786 between the T=0.21 and T=0.32 residuals is quoted without an uncertainty or a null-hypothesis test against uncorrelated noise of the same amplitude. A bootstrap or shuffle estimate would make the “same family of profiles” statement more quantitative.

Circularity Check

0 steps flagged

No significant circularity: residual is an operational leftover after nuisance scale-plus-offset fits and matched controls, not a prediction forced by construction.

full rationale

The paper reports an observational difference analysis, not a first-principles derivation or model prediction. For each train pair it defines R(q) = I_cys(q) - a I_water(q) - b with a,b fitted as nuisance parameters, then forms the control-corrected residual ΔR(q) = R_cys-water(q) - R_water-water(q). The claimed signal is whatever structure remains after these subtractions; a and b are removed rather than used to generate a forecast of the residual shape or amplitude. Integrated observables A+ and A- simply integrate the already-computed residual over fixed windows chosen for reproducibility; this is ordinary post-selection of display regions, not a fit that tautologically produces the reported significance. Convergence and block-averaging tests check 1/√N behaviour of the same residual ensemble and do not feed the residual definition. References are to instrument and SAXS literature with no load-bearing self-citation of a uniqueness theorem or ansatz that forces the residual. The analysis is therefore self-contained against its own operational definitions; any remaining scientific risk concerns experimental matching of controls, not circular reduction of claim to input.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 0 invented entities

The central claim rests on standard SAXS/XFEL measurement assumptions plus a few analysis choices (per-train linear model, water–water as non-sample residual proxy, fixed integration windows, fixed N=200). No new physical entities are postulated. Free parameters are nuisance fit coefficients and analysis-window choices, not a global theory fit that defines the residual.

free parameters (4)
  • per-train scale a and offset b
    Least-squares fitted independently for every cysteine–water (and water–water) train pair; remove detector-wide multiplicative/additive differences before residual analysis.
  • A+ integration window 0.05–0.15 Å⁻¹
    Chosen where positive residual lobe is most reproducible; defines the scalar observable used for transmission dependence and significance.
  • A− integration window 0.16–0.30 Å⁻¹
    Chosen for the negative residual lobe; intermediate q excluded by hand from integration.
  • N = 200 train pairs per transmission
    Fixed ensemble size for comparable precision across T settings; convergence tested on subsets up to this N.
axioms (4)
  • domain assumption A detector-wide scale-plus-offset model captures the dominant common-mode intensity differences between matched train profiles.
    Section III defines Icys(q)=a Iwater(q)+b and treats residual structure beyond a,b as the scientific observable.
  • domain assumption Transmission-matched water–water residuals estimate non-sample reproducibility limits under the same analysis.
    Sections III–IV define ΔR = R_cys−water − R_water−water as the central observable; scientific conclusions require this proxy to be adequate.
  • domain assumption After common-mode removal, independent train-pair residuals behave as approximately statistically independent samples (uncertainty ~1/√N).
    Section VI uses block averaging against 1/√N as evidence that high-repetition-rate averaging recovers expected sensitivity.
  • ad hoc to paper Only transmission settings with temporally close cysteine/water runs and available water–water controls are valid for primary fluence analysis.
    Section IV excludes T=0.046 and 1 M data for incomplete experimental symmetry; this selection defines the dataset supporting the claim.

pith-pipeline@v1.1.0-grok45 · 14916 in / 3272 out tokens · 32974 ms · 2026-07-14T12:24:51.437054+00:00 · methodology

0 comments
read the original abstract

We present a train-resolved SAXS methodology for recovering weak scattering signals from high-repetition-rate XFEL datasets and apply it to aqueous L-cysteine solutions measured at the European XFEL. Independent scale-plus-offset fitting was performed for matched cysteine and water train pairs, followed by subtraction of transmission-matched water--water controls. The 0.5 M dataset reveals a reproducible sign-changing residual SAXS signal that increases with incident XFEL transmission and remains after removal of detector-wide scaling, additive offsets, and matched water--water control residuals. Convergence and block-averaging analyses show that the residual emerges progressively as independent train pairs are accumulated and exhibits uncertainty scaling close to the expected inverse square-root dependence on N. These results establish a statistically robust transmission-dependent residual SAXS contribution whose microscopic origin remains unresolved, while demonstrating that train-resolved observables combined with matched controls can substantially improve sensitivity to weak scattering signals in high-repetition-rate XFEL experiments.

Figures

Figures reproduced from arXiv: 2607.10349 by Angelo Beratto-Ramos, Asier Garc\'ia, Biel Serrat, Carles Serrat, Chan Kim, Egor Sobolev, Huijong Han, Joana Valerio, Johan Bielecki, Katerina Doerner, Majed Chergui, Mohammad Vakili, Sara Hern\'andez, Tokushi Sato.

Figure 1
Figure 1. Figure 1: FIG. 1. Train-averaged residual structure for the 0.5 M cysteine–water comparisons after inde [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Control-subtracted residual SAXS structure for the 0.5 M cysteine dataset. The upper [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Convergence of the control-subtracted residual profiles as the number of train pairs in [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Block-averaging statistics of the integrated residual observables as a function of block [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

6 extracted references

  1. [1]

    Structural Modeling Using Solution Small-Angle X-ray Scattering (SAXS) , journal =

    Gr. Structural Modeling Using Solution Small-Angle X-ray Scattering (SAXS) , journal =. 2020 , volume =

    2020

  2. [2]

    and Round, Adam and Mertens, Haydyn D

    Blanchet, Clement E. and Round, Adam and Mertens, Haydyn D. T. and Ayyer, Kartik and Graewert, Melissa and Awel, Salah and Franke, Daniel and D. Form factor determination of biological molecules with X-ray free electron laser small-angle scattering (XFEL-SAS) , journal =. 2023 , volume =

    2023

  3. [3]

    Journal of Synchrotron Radiation , year =

    Allahgholi, Aschkan and others , title =. Journal of Synchrotron Radiation , year =

  4. [4]

    Megapixels @ Megahertz -- The AGIPD high-speed cameras for the European XFEL , journal =

    Allahgholi, Aschkan and Becker, Julian and Delfs, Annette and Dinapoli, Roberto and G. Megapixels @ Megahertz -- The AGIPD high-speed cameras for the European XFEL , journal =. 2019 , volume =

    2019

  5. [5]

    and others , title =

    Mancuso, Adrian P. and others , title =. Journal of Synchrotron Radiation , year =

  6. [6]

    3D Printed Devices and Infrastructure for Liquid Sample Delivery at the European XFEL , journal =

    Vakili, Mohammad and Bielecki, Johan and Kno. 3D Printed Devices and Infrastructure for Liquid Sample Delivery at the European XFEL , journal =. 2022 , volume =

    2022