{"id":"dcfb621f-a42c-4637-b6bf-58c59205f172","arxiv_id":"2412.19686","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"In two simulated tissue models, T1 cell rearrangements are Poisson-like in time but spatially biased by cell shape, and this bias is proposed to steer cells into directed migration.","lead":"This paper uses computer simulations of cell sheets to show that T1 neighbor-swapping events, although random in time, cluster unevenly around each cell and may push cells in particular directions. The findings offer a mechanical explanation for collective tissue movement, relevant to wound healing, development, and cancer invasion.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The migration/flow claim rests on orientation histograms, not on measured cell displacements; the causal direction between T1 asymmetry and motion is also untested.","rationale":"The paper's statistical characterizations of T1 events, waiting times, and shape-index changes are carefully presented and internally consistent. However, the headline assertion is about emergent directed migration and coherent tissue-scale flow. The only evidence offered is the non-uniformity of vertex-pair orientation histograms; the step from that geometric bias to net cell displacement is assumed, not demonstrated. The reader's weakest assumption correctly identified this gap, and it is load-bearing because the abstract and title assert migration as the outcome. I would strengthen the concern by noting the causal direction is also untested: in the MPF model, the very activity that drives T1s also drives cell motion, so the observed spatial patterning could be a byproduct of existing motion rather than its cause. This does not make the statistical results wrong, but it does make the central claim conditional. A direct measurement of scenario-conditioned cell displacements from the existing simulations would settle the issue, so the conditional verdict is appropriate and no further adjustment is needed.","tokens_in":14569,"tokens_out":3733,"duration_ms":43424,"concrete_test":"Compute in the MPF model the center-of-mass displacement Δr_i = r_i(t^+_{Tn}) - r_i(t^-_{Tm}) for every cell in each of the four successive-T1 scenarios, and compare the mean Δr_i direction with the orientation bias implied by Figs. 5f–i. The claim survives only if the scenario-conditioned mean displacement is significantly nonzero and aligned with the predicted direction; if the distribution of Δr_i is isotropic or anti-aligned, the geometric asymmetry does not cause directed migration.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section VI reports non-uniform relative-orientation histograms of vertices created in a first T1 and annihilated in a second T1 (Figs. 5f–i) and infers 'effective directed migration' from their non-uniformity. No cell-center displacement, velocity autocorrelation, mean-squared displacement, or tissue-scale flow field is presented anywhere in the paper. The abstract's claim that these asymmetric patterns 'promote directed cell migration, and form the backbone for coherent flow patterns at tissue scales' therefore goes beyond the data. The gap is not merely a missing diagnostic: in the MPF model T1s are triggered by the same self-propulsion activity (Eqs. A4–A5) that moves cells, so the observed spatial bias could be a consequence of a cell's existing directed motion selecting where rearrangements occur rather than the rearrangements generating the motion. The paper's own hedged wording in Section VI ('suggest', 'could lead') and in the Conclusions ('potentially') does not bridge this. The orientation histograms may be correct and still not produce net migration if, e.g., the two T1s' quadrupolar flows partially cancel or the cell's own activity dominates its trajectory.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the spatiotemporal statistics of T1 transitions in two complementary models of epithelial monolayers: a multi-phase field (MPF) model with deformable cells and an active elastic solid (AES) model with point-like cells. The authors report three main results: (i) successive T1 transitions involving a given cell follow an exponential waiting-time distribution, i.e., a Poisson process in time; (ii) T1 events are spatially correlated, as quantified by radial distribution functions of T1 epicenters; and (iii) cell shape responds asymmetrically to T1 events, with loser cells relaxing and gainer cells elongating. The central claim is that non-uniform spatial distributions of vertices created and annihilated in successive T1s (Figures 5f-5i) lead to directed cell migration and, through chaining of T1s, to coherent tissue-scale flow patterns. The paper presents these statistics from simulations but does not directly measure cell displacement or flow.","tokens_in":14800,"tokens_out":3192,"duration_ms":32353,"significance":"If the causal link between the observed T1 asymmetry and net cell migration is established, the paper would provide a concrete single-cell topological mechanism for collective migration in epithelia, connecting local junction remodelling to tissue-scale flow. The strengths are that the exponential waiting-time distributions, the radial distribution functions, and the shape-index statistics are computed directly from simulations, and the authors include robustness checks for the alpha = 0 case and for varying activity and deformability. The two-model approach (MPF and AES) strengthens the generality of the temporal Poisson statistics. However, the headline claim about directed migration and coherent flow currently rests on geometric orientation histograms rather than on measured cell displacements or flow fields, so the paper's significance is conditional on additional direct evidence.","major_comments":[{"comment":"The central claim that asymmetric spatial patterns of successive T1 transitions 'promote directed cell migration, and form the backbone for coherent flow patterns at tissue scales' is not directly supported by the data shown. The paper never presents cell-center displacements, velocity autocorrelations, mean-squared displacements, or tissue-scale velocity fields. The non-uniform orientation histograms of vertices in Figures 5f-5i are geometric biases of where rearrangements occur, but they do not by themselves establish net cell motion. Moreover, in the MPF model, T1 transitions are triggered by the same self-propelled activity that moves cells (Eqs. A1, A4-A5), so the observed spatial bias could be a consequence of an already-directed motion selecting where T1s occur rather than the T1s generating the motion. A concrete test would be to compute the mean-squared displacement or velocity autocorrelation of cells conditioned on the four scenarios in Figure 5, or to measure the tissue-scale velocity field and its correlation with the T1 orientation histograms.","section":"Section VI, Figs. 5f-5i; Abstract and Conclusions"},{"comment":"The statement that 'these distributions remain similar even if activity and deformability are varied suggesting a universal emergent behaviour' is not supported by any displayed figure or quantitative comparison. The supplemental material shows only the alpha = 0 case (Supp. Fig. 2), not variations of v0 and Ca for the relative-orientation histograms. To make the universality claim, the authors should either include the histograms for varied v0 and Ca or provide a quantitative similarity measure (e.g., histogram overlap or a distance metric).","section":"Section VI, paragraph after Fig. 5k"}],"minor_comments":[{"comment":"In the paragraph discussing Figure 5, the text refers to 'Figure 5g' for both the loser-gainer and the gainer-loser scenarios; the second reference should likely be Figure 5h.","section":"Section VI"},{"comment":"The word 'emply' should be 'employ' in the sentence 'we emply two complementary modeling approaches'.","section":"Introduction"},{"comment":"The neighbor criterion is described once as 'cells i and j are considered neighbours if and only if they are closest in distance to their midpoint than any other third cell' and then given explicitly in equations; the verbal phrasing is ambiguous and should be reworded to match the exact mathematical condition.","section":"Appendix B"},{"comment":"The claim that waiting times are 'exponentially distributed' is made by eye; providing the mean rates and, if possible, a goodness-of-fit or a log-linear plot would strengthen the statistical statement.","section":"Section III, Fig. 2"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the scope of physics.bio-ph and presents solid simulation statistics. The main reservation is that the headline claim about directed migration and tissue-scale flow is not measured directly; this is a load-bearing gap that is fixable with additional displacement or flow analysis. I recommend major revision rather than rejection because the underlying orientation histograms and shape statistics are likely correct, and the missing direct evidence is well within the scope of the manuscript to supply."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What to know: the new thing here is the four-scenario decomposition of successive T1 transitions (loser-loser, loser-gainer, gainer-loser, gainer-gainer) and the histograms of relative orientation between created and annihilated vertices. That is a genuinely useful way to characterize the spatial structure of T1 sequences, and the simulations are competently done. The problem is the headline: the abstract says these asymmetric patterns 'promote directed cell migration, and form the backbone for coherent flow patterns,' but the paper never measures migration or flow.\n\nCredit where due: the waiting-time distributions are clean exponentials with activity-dependent rate, and the shape-index evolution separating loser from gainer cells (Fig. 4a) is a nice, clear result. The robustness checks — α=0, varying Ca and v0 — are reassuring, and the AES model provides a consistency check for the Poissonian waiting times. The statistics in Figs. 5f-5i are computed directly from the simulation trajectories, not fitted, so the empirical content is solid.\n\nWhere it gets soft: the causal claim. There is no velocity autocorrelation, no mean-squared displacement of cell centers, no flow-field analysis anywhere in the paper. The non-uniform orientation histograms are interpreted in Section VI as suggesting 'an effective directed migration,' and the conclusions escalate that to 'allowing cells to move directionally... potentially forming flow structures.' The stress-test note has it right: the same self-propulsion activity that triggers T1s also moves the cell, so the spatial bias could be a consequence of existing directed motion rather than its cause. The paper's own hedged language doesn't bridge that. A couple of histograms could be correct and still produce no net migration if opposing biases cancel or the cell's own activity dominates its trajectory.\n\nMinor issues: no code or data release, which will slow replication. And the AES model is used only for waiting times and g(r); all the shape-dependent analysis comes from MPF, so the 'two complementary models' framing oversells the AES contribution slightly. Neither of these changes my overall read.\n\nBottom line: the statistical core deserves a serious referee. The paper should go to review, but the migration/flow claim needs to be either directly measured (MSD, velocity autocorrelation, or tissue-scale flow field) or explicitly softened to a hypothesis. If the authors do that, this becomes a solid contribution to the T1-transition literature.\n\nWho this is for: anyone working on cell rearrangements, shape feedback, and collective migration in epithelia. I'd bring it to a reading group, and I'd cite the four-scenario histograms if I needed a spatial characterization of successive T1s.","headline":"A clean new statistical decomposition of successive T1 transitions undermined by a migration claim that no measured cell displacement supports.","tokens_in":15327,"tokens_out":3140,"would_cite":true,"duration_ms":30894,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Although the four-cell neighbor swaps called T1 transitions occur randomly in time for any one cell, the paper shows that their spatial distribution is biased by cell shape, and that this asymmetry gives cells a directional tendency that…","keywords":["T1 transitions","epithelial monolayers","collective cell migration","multi-phase field model","active elastic solid model","cell shape index","tissue fluidization","topological rearrangements"],"falsifier":"Track the center of mass of each cell through a loser-gainer or gainer-loser pair of successive T1s in the multi-phase field model and average the displacement conditioned on the scenario: if the mean displacement is zero, or points opposite to the direction implied by the relative-orientation histograms, then the non-uniform spatial distribution does not by itself produce directed migration.","tokens_in":14353,"feed_emoji":"🔄","tokens_out":7567,"duration_ms":395726,"temperature":0.7,"pith_summary":"T1 transitions, localized four-cell rearrangements that swap neighbors in an epithelial monolayer, look random in time but are spatially organized, and this paper argues that the spatial organization is set by cell shape. Using two complementary simulations, a multi-phase field model that resolves cell contours and an active elastic solid model of self-propelled cells connected by springs, the authors show that losing a neighbor relaxes a cell's shape while gaining a neighbor elongates it. They then track the vertices created and destroyed in successive T1s and find that their locations are non-uniformly distributed around the cell in four distinct loss/gain scenarios. The paper proposes that this geometric bias is what turns isolated topological rearrangements into directed cell migration and, ultimately, into coherent tissue-scale flows.","feed_headline":"Neighbor-swap geometry steers cell migration","feed_subtitle":"Rearrangements are random in time but biased in space, and the bias can build tissue-scale flow.","key_machinery":"The central object is the T1 transition, a four-cell topological rearrangement in which the junction between two cells shrinks away while two other cells form a new junction roughly perpendicular to the old one. The argument is carried by the cell shape index $p=\\mathrm{Perimeter}/\\sqrt{\\mathrm{Area}}$, a measure of elongation, and by the relative orientation $\\xi$ between vertices created in the first of two successive T1s and vertices destroyed in the second, measured around the cell's center of mass. Histograms of $\\xi$ for the four successive-T1 scenarios are the load-bearing quantity: a non-uniform histogram means the next rearrangement tends to occur on a particular side of the cell, which is the microscopic mechanism proposed for directed migration. The waiting-time distributions and radial distribution functions support the temporal randomness and spatial correlation of the events.","core_discovery":"The paper's central claim is that T1 transitions, random in time for a single cell, are spatially correlated across the tissue and biased by cell shape. Cells that lose neighbors relax toward a rounder shape, while cells that gain neighbors elongate; the paper documents this dual effect with the shape index $p=\\mathrm{Perimeter}/\\sqrt{\\mathrm{Area}}$. For each of the four ways a cell can participate in two successive T1s, losing twice, losing then gaining, gaining then losing, or gaining twice, the authors measure the relative orientation between vertices created by the first T1 and vertices annihilated by the second, and find non-uniform histograms. These non-uniform distributions are the evidence that successive T1s happen preferentially on one side of a cell, giving the cell a directional bias. The paper concludes that such biases chain together into coherent flow patterns at tissue scales.","pith_inferences":["Editorial inference: the same relative-orientation histograms could be gathered from live monolayers by tracking tri-cellular vertices; positive results would make the geometric bias a directly observable tissue property rather than a simulation-only statistic.","Editorial inference: if shape deformation is the driver, then suppressing deformation, for example by raising junctional tension or lowering deformability in the model, should weaken the orientation bias and the associated flow coherence; this is a parameter sweep the paper does not report.","Editorial inference: because a gainer-gainer chain implies a neighboring loser-loser chain, the cell-level directional bias may be compensated at the level of neighbor pairs, possibly producing local shear rather than pure translation; quantifying that compensation would test how the coherent flow actually forms."],"forward_implications":["T1 events can be described as temporally random with an activity-dependent rate, while their spatial distribution is correlated over several cell diameters.","Neighbor loss relaxes a cell and neighbor gain stretches it, so the same rearrangement has opposite mechanical effects on the cells that shed a contact and those that form a new one.","Successive T1s do not occur uniformly around a cell; the distributions differ between lose-gain and gain-lose, so the order of topology changes matters for the cell's directional response.","The orientation bias persists when the model's activity vector no longer aligns with cell elongation, pointing to cell shape deformation, not polarity alignment, as the source of the spatial organization.","Chains of these biased rearrangements can organize into coherent, tissue-scale flows, linking single-cell topology changes to collective migration."],"supporting_citations":[{"why":"Supplies the multi-phase field model and earlier robust statistical properties of T1 transitions, including shape-index evolution and waiting-time behavior.","marker":"[8]"},{"why":"Provides the model formulation used here and the earlier finding that T1s generate quadrupolar flows and enhance relative dispersion, which the present paper extends to spatial patterning.","marker":"[9]"},{"why":"Establishes the earlier multi-phase field framework for topological and geometrical quantities in active cellular structures.","marker":"[17]"},{"why":"Introduces the active elastic solid model of self-propelled, spring-connected particles that the paper extends with neighbor exchanges.","marker":"[23]"},{"why":"Supplies the geometric nearest-neighbor criterion used to trigger T1 transitions in the active elastic solid model.","marker":"[25]"}],"fun_headline_variants":["Neighbor swaps are random, yet their bias steers cells","Cell shape changes turn random swaps into directed flow","Asymmetric cell rearrangements produce tissue-scale drift","T1 transitions: random in time, biased in space, directive","How random cell swaps build coherent migration"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the non-uniform spatial distribution of successive T1s actually produces net cell displacement; the simulations measure the geometry of the T1 locations but do not directly measure cell migration or tissue flow.","fun_headline_variants_meta":{"raw":{"variants":["Neighbor swaps are random, yet their bias steers cells","Cell shape changes turn random swaps into directed flow","Asymmetric cell rearrangements produce tissue-scale drift","T1 transitions: random in time, biased in space, directive","How random cell swaps build coherent migration"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000571,"raw_usage":{"total_tokens":2649,"prompt_tokens":840,"completion_tokens":1809,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":456,"completion_tokens_details":{"reasoning_tokens":1733}},"tokens_in":456,"tokens_out":1809,"duration_ms":14409,"temperature":1.0,"reasoning_tokens":1733,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:58:06.702777+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track the center of mass of each cell through a loser-gainer or gainer-loser pair of successive T1s in the multi-phase field model and average the displacement conditioned on the scenario: if the mean displacement is zero, or points opposite to the direction implied by the relative-orientation histograms, then the non-uniform spatial distribution does not by itself produce directed migration.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the multi-phase field model and earlier robust statistical properties of T1 transitions, including shape-index evolution and waiting-time behavior."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the model formulation used here and the earlier finding that T1s generate quadrupolar flows and enhance relative dispersion, which the present paper extends to spatial patterning."},{"cited_title":"Wenzel, S","cited_arxiv_id":null,"evidence_quote":"Establishes the earlier multi-phase field framework for topological and geometrical quantities in active cellular structures."},{"cited_title":"Ferrante, A","cited_arxiv_id":null,"evidence_quote":"Introduces the active elastic solid model of self-propelled, spring-connected particles that the paper extends with neighbor exchanges."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the geometric nearest-neighbor criterion used to trigger T1 transitions in the active elastic solid model."}],"review_version":1}