{"id":"9ccdae58-8528-45e3-8633-5d8a0f08e377","arxiv_id":"2507.18887","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"In 2D gastruloids, colony size determines pattern symmetry, mesodermal pattern area follows a near-linear power-law scaling with colony area, and growth switches from power-law expansion to exponential arrest, captured by a reactive-boundary Turing model.","lead":"Small flat colonies of stem cells form off-center mesodermal patterns while large ones stay symmetric and centered, and the patterned area grows with colony size as a power law. The paper proposes a Turing-style activator-repressor model with self-adjusting boundaries to explain these size-dependent behaviors and the two-phase growth that ends in arrest.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed quantitative match between experimental (ε=1.92±0.19) and simulated (ε=1.44±0.06) final-pattern size scaling is not supported as reported, and the simulated ε is inconsistent with the model's own steady-state scaling exponent α≈0.9.","rationale":"I focused on the internal consistency of the model's scaling exponents because it is a load-bearing, checkable component of the central claim, independent of unmeasured physical parameters. The paper asserts a quantitative match that appears to be contradicted by its own numbers, which is more directly falsifying than the unit-conversion fragility identified by the reader. Nevertheless, the empirical findings (size-dependent symmetry-breaking, scaling, biphasic growth) are substantial and may stand regardless; the concern is specifically about the model's claimed quantitative reproduction. A reanalysis would settle the issue, so the conditional verdict remains appropriate.","tokens_in":11437,"tokens_out":10033,"duration_ms":102602,"concrete_test":"Reanalyze the simulation data used for Fig. 3b and Fig. 2d with identical definitions of pattern area and fitting ranges; verify whether ε=2α holds. Also re-extract the experimental Smax_p vs R exponent from the full dataset including R≈50 µm colonies, reporting the 95% CI. If the simulated ε is actually ≈1.8, then the published 1.44 is an error and the comparison to the experimental 1.92 may be valid; if ε remains 1.44, then the model's scaling laws are internally inconsistent and the claimed match to the experimental α=0.93±0.13 fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the Turing model reproduces the experimental scaling behavior is undermined by an internal inconsistency in the reported exponents. Figure 3b reports the simulated maximum pattern area Smax_p scales with radius as ε=1.44±0.06. Figure 2d reports the simulated steady-state scaling of pattern area with colony area as α≈0.9 (for δ_B∈[2,4]), which implies Smax_p ~ R^{2α} ≈ R^{1.8}. These two model-derived exponents cannot both be correct. Moreover, the experimental Smax_p scaling is measured as ε=1.92±0.19, which differs from the cited simulated 1.44 by about 2.4 standard errors (p≈0.016), so the claim that the experiment 'matches' the simulation is not supported. Because this comparison uses unitless exponents, it does not depend on the fitted λ=50 µm or k_A=0.14 h⁻¹ conversion, making it a direct, robust falsifier of the quantitative agreement. If the model cannot reproduce the size-dependence of the final pattern area, the broader claim that it reproduces the observed scaling and critical growth transitions loses its quantitative foundation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript combines high-throughput fixed and live imaging of 2D adherent mouse PSC gastruloids on micropatterned substrates with a two-component Turing activator-repressor model (Eqs. 1-4) with reactive boundary fluxes. It reports three central findings: (i) colony size controls the symmetry of the Bra+ mesodermal pattern, with small colonies (R≈50-150 µm) breaking symmetry (χ≈0.6) and large colonies (R>200 µm) remaining centro-symmetric (χ≈0.2); (ii) the pattern area obeys a power-law scaling Sp ~ Sc^α with α=0.93±0.13 in fixed samples (N=15,305) and α=0.95±0.03 in live imaging (N=520), independent of seeding density; and (iii) pattern growth is biphasic, with an early power-law expansion (exponent γ) followed by exponential arrest (rate β), with a transition at t†≈22-27 h that the authors interpret as a dynamical phase transition. The model is argued to reproduce these behaviors when the morphogen decay length is λ=50 µm, the repressor penetration length is δ_B∈[2,4], and the degradation rate is kA=0.14 h⁻¹. The manuscript claims a unified self-organizing mechanism for symmetry-breaking, scaling, and growth control in organoids.","tokens_in":11719,"tokens_out":9496,"duration_ms":93676,"significance":"If the quantitative claims held, the paper would be a significant contribution to developmental biophysics and organoid engineering: it uses an unusually large dataset, provides time-resolved scaling and growth exponents, and demonstrates that a minimal reaction-diffusion model with reactive boundaries can qualitatively connect symmetry-breaking, size scaling, and growth arrest. The distinction between the morphogen concentration field and the morphogen production region as the instructive patterning cue is an interesting and testable hypothesis. The strengths include the explicit reporting of sample sizes and fit statistics, the master-curve collapse analysis, and the clear operational definitions of the symmetry score and scaling exponents. However, the manuscript currently contains a load-bearing internal inconsistency in the simulated exponents and a calibration procedure that overlaps with the validation data. These issues prevent the current version from establishing the claimed quantitative match.","major_comments":[{"comment":"The two simulated scaling exponents reported for the same model are mutually inconsistent. Because the colony is a disc, Sc = πR², so Eq. (6), Sp∼Sc^α, is equivalent to Sp∼R^{2α}. The text in §3 states that simulations with δ_B∈[2,4] reach steady-state values α≈0.9 (Fig. 2d), which implies Smax_p∼R^{1.8}. Fig. 3b's inset instead reports Smax_p∼R^ε with ε=1.44±0.06. Since both exponents describe the final pattern area as a function of colony size in the same simulations, they cannot both be correct. This is not merely a presentational slip: the experimental maximum-area exponent is ε=1.92±0.19 (Fig. 3f inset), and the difference from the simulated 1.44 is about 2.4 combined standard errors, so the stated 'match' between experiment and simulation is not supported. The authors should recompute both exponents from the same simulation outputs and either correct ε or explicitly explain why Smax_p and the steady-state S_p plotted in Fig. 2d are different observables.","section":"Fig. 2d vs Fig. 3b; §3–§4"},{"comment":"The model's agreement with experiment is weakened by the fact that the same data are used for calibration and validation. The decay length λ=50 µm is assumed to place the simulated symmetry transition at the observed colony sizes, δ_B is restricted to [2,4] explicitly to reproduce the experimental scaling exponent α≈0.9, and kA is fitted from the τ*–t* regression in Fig. 1j (R²=0.64, kA=0.14±0.12 h⁻¹). Consequently, the 'match' between the simulated and experimental α values in Fig. 2 is not an independent test of the model. To make the comparison informative, the authors should report the model's predicted α(δ_B, δ_A) before calibration with the experimental value, show the sensitivity of the predicted ε and β to the uncertainty in λ and kA, or provide an out-of-sample prediction (e.g., a perturbation experiment) that does not depend on the tuned parameters.","section":"§2–§3, calibration of λ, kA, and δ_B"},{"comment":"The biphasic growth law and the claim of exponential arrest are based on data that are censored at both ends. The text explicitly excludes R≈50 µm colonies from the scaling because they 'misalign with the overall scaling trend,' and restricts the live-imaging analysis to t<40 h because the Bra+ domain shrinks beyond that time. The exponential relaxation exponent β and the transition time t† are extracted from these truncated trajectories. Without a sensitivity analysis (e.g., refitting γ and β with the truncation point varied between 30 and 40 h, and with R=50 colonies either included or excluded), it is not clear whether the reported biphasic law and the 'critical' transition are robust features or artifacts of censoring. This matters because the central claim of a dynamical phase transition rests on the existence and sharpness of t†.","section":"§4, 'Power-law growth and critical arrest…', Fig. 3e–g"}],"minor_comments":[{"comment":"The symmetry scores for small and large colonies (χ=0.6±0.2 vs 0.2±0.2) are reported without a statistical test, confidence interval, or effect size; please add a formal comparison at least for the low-density condition.","section":"§1, Fig. 1e"},{"comment":"The text reports kA=0.14 h⁻¹ from a linear fit with R²=0.64, but the caption gives kA=0.14±0.12 h⁻¹; the dimensional conversion and all derived quantities (DA=0.10 µm²/s, β in h⁻¹) should be accompanied by propagated uncertainties.","section":"§2, Fig. 1j"},{"comment":"The text says the scaling spans 'two orders of magnitude in gastruloid size,' but radii from 50 to 400 µm correspond to a 64-fold range in area, not two orders of magnitude; the claim should be reworded or the relevant variable specified.","section":"§3, Fig. 2f"},{"comment":"The production domain S_s is defined by an integral of the production function P, but the text later refers to 'source region' and 'non-source region' without stating the threshold used to segment the simulated domain; please define the operational threshold.","section":"§2, Eq. (5)"},{"comment":"The use of 'dynamical phase transition' would benefit from a quantitative criterion; the current evidence rests on peaks in higher derivatives of the growth rate, which are not a standard order-parameter or scaling-collapse test. Consider softening the terminology or adding such a test.","section":"§4"}],"recommendation":"major_revision","confidential_remarks":"The main blocker is the internal inconsistency between the simulated ε and α values (Fig. 2d and Fig. 3b). Please ask the authors to provide the raw simulation outputs and re-derived exponents before the next round; if the corrected ε turns out to be approximately 1.8, the quantitative comparison with the experimental ε=1.92 becomes much more favorable, but as written the manuscript contains contradictory numbers."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a substantial empirical paper whose qualitative story—size controls symmetry, pattern area scales with colony area, and growth arrests in a biphasic way—is credible and worth a serious referee. The problem is the claimed quantitative match between model and experiment. It does not hold together as reported, and one internal inconsistency in the model's exponents looks like a genuine flaw, not a stylistic quibble.\n\nWhat is new: 2D adherent gastruloids show spontaneous symmetry-breaking in small colonies (R≈50–150 µm) and centro-symmetric patterns in larger ones; the Bra+ domain scales with colony area as Sc^0.93±0.13 across ~15,000 fixed and ~520 live colonies, apparently robust to density; growth is biphasic with a transition at 22–27 h. Those are solid empirical findings with unusually large N. The reactive-boundary Turing model is a reasonable minimal framework and does reproduce the qualitative phase behavior over a sensible parameter range.\n\nThe soft spots: the model is calibrated to the same data in several places—λ=50 µm is assumed to place the symmetry transition, kA comes from a fit with R²=0.64 and error ±0.12 h⁻¹, and δ_B is selected in [2,4] to match the experimental α. The stress-test point is sharper and independent of those calibrations: Figure 3b reports simulated maximum pattern area scaling Smax_p~R^1.44, while Figure 2d reports a steady-state α≈0.9 for δ_B=4 (and 0.98 for δ_B=2). Since Sc~R^2, α≈0.9 implies Smax_p~R^1.8. Both cannot be right. The experimental maximum-pattern-area exponent is 1.92±0.19, which is about 2.4 standard errors away from 1.44. So the paper's specific claim that the experiment matches the simulation on final pattern size is not supported. Because this is a unitless comparison, it does not depend on the λ/kA conversion and can't be blamed on the unit mapping. Separately, R≈50 µm colonies are excluded from the scaling analysis and experimental growth curves are truncated at ~40 h; each is arguably defensible, but the treatment needs to be explicit and justified. The 'critical transition' is best framed as a modeled crossover between two fitted regimes, not a demonstrated phase transition.\n\nIf the model cannot reproduce the size dependence of the final pattern area, the broader claim that it explains the observed scaling and growth arrest loses its quantitative foundation. The empirical results remain valuable, and the model is a productive starting point, but the manuscript needs major revision before the quantitative claims can be accepted.\n\nMy recommendation: send it to peer review. The data deserve serious referee time, and the inconsistency is fixable—but for now, conditional acceptance would be premature.","headline":"A data-rich organoid study with a credible qualitative story and a real internal inconsistency in the model's own scaling exponents—worth refereeing, but the quantitative match as reported does not hold.","tokens_in":12313,"tokens_out":2177,"would_cite":true,"duration_ms":24269,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["92C15","92C42","35Q92"],"pacs":[],"model":"deepseek-v4-flash","headline":"A minimal two-component Turing system with reactive boundaries explains size-dependent symmetry-breaking, power-law scaling, and biphasic growth arrest in 2D gastruloids.","keywords":["self-organization","symmetry-breaking","Turing pattern","reaction-diffusion system","gastruloid","scaling law","critical transition","stem-cell organoid"],"falsifier":"Measure the actual decay length and degradation rate of the activator/repressor morphogens in this system—for example by fluorescently tagged ligands or local photobleaching—and check whether $\\lambda \\approx 50$ µm and $k_A \\approx 0.14$ h$^{-1}$; if either differs substantially, the predicted symmetry transition at $R \\approx 150{-}200$ µm, the scaled transition time $t^\\dagger$, and the derived $D_A = 0.10$ µm$^2$/s would not hold. A second falsifier: block boundary reactivity, for example by changing the micropattern chemistry so that $\\sigma_i \\to 0$, and observe whether small colonies that currently break symmetry instead stay centro-symmetric.","tokens_in":11196,"feed_emoji":"🧫","tokens_out":8948,"duration_ms":84513,"temperature":0.7,"pith_summary":"2D adherent gastruloids—flat stem-cell colonies grown on micropatterned discs—organize their mesodermal (Brachyury+) domain according to colony size: small colonies (radius 50–150 µm) break symmetry into a bilateral pattern, while large colonies (radius >200 µm) stay centro-symmetric. The domain area follows a power law in colony area, with exponent $\\alpha = 0.93 \\pm 0.13$ ($0.95 \\pm 0.03$ in live imaging), across a 20-fold density range. Time-lapse growth of the pattern is biphasic: an early power-law expansion gives way to exponential arrest, with the switch at $t^\\dagger = 22{-}27$ h, which the authors read as a dynamical phase transition. The paper argues that all three phenomena come from one mechanism: a two-component Turing activator–repressor system with concentration-dependent reactive boundaries, where dimensionless radius and diffusion-penetration lengths set the regimes. If right, this gives a minimal physical explanation for how tissues sense size and self-arrest growth without external gradients.","feed_headline":"Small 2D colonies break symmetry; large ones don't","feed_subtitle":"A Turing model with reactive boundaries reproduces scaling and a biphasic growth arrest across 15,000 colonies.","key_machinery":"The load-bearing object is a minimal two-component reaction–diffusion system on a disc: an activator $A$ and repressor $B$ with equations $\\partial_t A = \\nabla\\cdot(D_A\\nabla A) - k_A A + \\nu_A P$ and $\\partial_t B = \\nabla\\cdot(D_B\\nabla B) - k_B B + \\nu_B P$, where $P = A^h/(A^h+B^h)$ is a sharp switch. The twist is the boundary condition: instead of fixed concentrations or zero flux, fluxes are reactive, $D_A\\nabla A\\cdot\\hat{n} = \\sigma_A A$ and $D_B\\nabla B\\cdot\\hat{n} = \\sigma_B B$, so the edge feeds back on the bulk. In dimensionless form the system is controlled by $\\hat{R} = R/\\lambda$ and $\\hat{\\delta}_i = D_i/(\\sigma_i \\lambda)$; these two ratios determine whether the colony polarizes and how the pattern area scales with colony area. The machinery works by letting the production domain itself grow, scale, and arrest, coupling patterning to size and to the boundary without an imposed gradient.","core_discovery":"On the paper's own terms, the central discovery is that size, not external cues, determines the mode and extent of mesodermal patterning in 2D gastruloids. A symmetry score $\\chi$ separates small colonies ($\\chi \\approx 0.6$, bilateral) from large ones ($\\chi \\approx 0.2$, centro-symmetric), and the same boundary-feedback model that produces this phase diagram also produces the observed power-law scaling $S_p \\sim S_c^{\\alpha}$ and the biphasic growth law with a size-invariant transition time. The model's dimensionless numbers—normalized radius $\\hat{R} = R/\\lambda$ and normalized penetration lengths $\\hat{\\delta}_i = D_i/(\\sigma_i \\lambda)$—predict that crossing a size threshold moves the system from symmetry-breaking to centro-symmetric patterns, that the repressor's penetration depth $\\hat{\\delta}_B$ sets the scaling exponent ($\\alpha \\approx 0.9$ for $\\hat{\\delta}_B \\in [2,4]$), and that pattern area at arrest scales as $S_p^{\\max} \\sim R^{\\epsilon}$ with $\\epsilon = 1.92 \\pm 0.19$ experimentally versus $1.44 \\pm 0.06$ in simulations. The authors further argue that the fate domain is set by the morphogen production region, $P = A^h/(A^h+B^h)$, rather than by a fixed threshold on concentration, so the source itself is the dynamic patterning agent.","pith_inferences":["Treating the biphasic growth as a critical transition suggests a testable signature: near $t^\\dagger$, the variance of growth rates across colonies or the response to small perturbations should peak if the system is genuinely critical; the paper's derivative analysis hints at this but does not measure susceptibility directly.","The same reactive-boundary Turing mechanism may explain size-invariant scaling in other organoid systems, since the dimensionless ratios $R/\\lambda$ and $D/(\\sigma\\lambda)$ are generic; one prediction is that changing substrate adhesiveness ($\\sigma_i$) alone should switch the symmetry regime at fixed $R$.","Because the unit conversion rests on assumed $\\lambda = 50$ µm and fitted $k_A = 0.14$ h$^{-1}$, the reported dimensional diffusion coefficient $D_A = 0.10$ µm$^2$/s should be treated as a model estimate, not a measurement; direct measurement of morphogen decay length would decide how much of the quantitative agreement is real.","The model's identification of the production region as the instructive signal suggests a new experimental handle: inducible ligand secretion or ligand biosensors could distinguish source-based from threshold-based fate specification in the same colony."],"forward_implications":["Spontaneous axis formation does not require 3D tissue organization; 2D colonies are sufficient when they are small enough.","Cells collectively estimate colony size through a diffusible morphogen whose penetration depth relative to colony radius sets the pattern, giving a mechanistic basis for positional information.","Growth arrest is self-organized rather than timed externally: the same feedback that forms the pattern slows its own expansion, with a common transition time $t^\\dagger \\approx 22{-}27$ h across colony sizes.","The scaling exponent $\\alpha$ is set by the repressor's normalized penetration depth $\\hat{\\delta}_B$, so perturbations that change $\\hat{\\delta}_B$ (for example, altering cell density or boundary reactivity) should move $\\alpha$ in a predictable direction.","The model predicts that the mesodermal fate domain tracks the morphogen production region, so visualizing production dynamics of BMP4/Wnt/Nodal-type ligands should show source-region expansion and arrest matching the Bra+ domain."],"supporting_citations":[{"why":"Defines the micropatterned colony system and supplies the anterior–posterior length scale used to choose radii from 50 to 400 µm.","marker":"[16]"},{"why":"Supplies the micropatterning method used to confine colony area and shape.","marker":"[19]"},{"why":"Earlier demonstration of Turing-type activator–repressor patterning in stem-cell colonies that this work extends with reactive boundaries.","marker":"[22]"},{"why":"Supplies the reaction–diffusion instability theory underlying the two-component model.","marker":"[28]"},{"why":"Provides the sharp-switch production term $P = A^h/(A^h+B^h)$ used in the model.","marker":"[31]"},{"why":"Motivates concentration-dependent reactive boundary fluxes for morphogen dynamics at the colony edge.","marker":"[26, 27]"},{"why":"Supports the idea that the morphogen production region, rather than a fixed concentration threshold, acts as the dynamic patterning source.","marker":"[33]"},{"why":"Provides reported morphogen diffusion length scales used to justify $\\lambda = 50$ µm.","marker":"[34]"}],"fun_headline_variants":["Size sets organoid symmetry: small break, large stay","Organoid size flips symmetry and growth mode","Critical scaling and symmetry switch by colony size","Biphasic growth and symmetry tied to colony size"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative match stands on the assumed morphogen decay length $\\lambda = 50$ µm and the fitted degradation rate $k_A = 0.14 \\pm 0.12$ h$^{-1}$, which convert the model's dimensionless radius and time into physical radius and time but are not independently measured.","fun_headline_variants_meta":{"raw":{"variants":["Size sets organoid symmetry: small break, large stay","Organoid size flips symmetry and growth mode","Critical scaling and symmetry switch by colony size","Biphasic growth and symmetry tied to colony size"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000604,"raw_usage":{"total_tokens":2891,"prompt_tokens":1091,"completion_tokens":1800,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":707,"completion_tokens_details":{"reasoning_tokens":1738}},"tokens_in":707,"tokens_out":1800,"duration_ms":16329,"temperature":1.0,"reasoning_tokens":1738,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:06:32.770342+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the actual decay length and degradation rate of the activator/repressor morphogens in this system—for example by fluorescently tagged ligands or local photobleaching—and check whether $\\lambda \\approx 50$ µm and $k_A \\approx 0.14$ h$^{-1}$; if either differs substantially, the predicted symmetry transition at $R \\approx 150{-}200$ µm, the scaled transition time $t^\\dagger$, and the derived $D_A = 0.10$ µm$^2$/s would not hold. A second falsifier: block boundary reactivity, for example by changing the micropattern chemistry so that $\\sigma_i \\to 0$, and observe whether small colonies that currently break symmetry instead stay centro-symmetric.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the micropatterned colony system and supplies the anterior–posterior length scale used to choose radii from 50 to 400 µm."},{"cited_title":"Warmflash, B","cited_arxiv_id":null,"evidence_quote":"Supplies the micropatterning method used to confine colony area and shape."},{"cited_title":"Tewary, J","cited_arxiv_id":null,"evidence_quote":"Earlier demonstration of Turing-type activator–repressor patterning in stem-cell colonies that this work extends with reactive boundaries."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the reaction–diffusion instability theory underlying the two-component model."},{"cited_title":"Werner, T","cited_arxiv_id":null,"evidence_quote":"Provides the sharp-switch production term $P = A^h/(A^h+B^h)$ used in the model."},{"cited_title":"Aguilar-Hidalgo, S","cited_arxiv_id":null,"evidence_quote":"Supports the idea that the morphogen production region, rather than a fixed concentration threshold, acts as the dynamic patterning source."},{"cited_title":"Kicheva, P","cited_arxiv_id":null,"evidence_quote":"Provides reported morphogen diffusion length scales used to justify $\\lambda = 50$ µm."}],"review_version":1}