{"id":"5af7e34d-a13e-48d2-8de4-8409f00ee0b1","arxiv_id":"2501.04827","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In epithelial monolayers, +1/2 topological defects move either tail-to-head or head-to-tail depending on whether energy is injected by cell-substrate tractions or intercellular stresses, and the driving force patterns exist before the defects form.","lead":"Experiments on sheets of kidney cells show that the same layer can contain topological defects in cell alignment moving in opposite directions at the same time. The direction is set by whether the energy driving the motion comes from forces between cells or from the cells pulling on their underlying surface.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Stress-reconstruction assumption is load-bearing: the 'stress-dominated' half of the claim depends on signs of PS and EΓ that may be artifacts of the passive linear constitutive model in monolayer stress microscopy.","rationale":"The reader's weakest assumption already identified the monolayer stress microscopy constitutive closure as the central risk, and I agree. The direct measurements (coexistence of tail-to-head and head-to-tail motions, traction-velocity angle, and counter-rotating vortices) support the observational core. What remains load-bearing is the partition of energy injection between stress and traction, which is inferred from a stress tensor reconstructed under a passive linear closure. Because the Clapeyron balance is an identity for any equilibrium-satisfying stress, it cannot independently confirm the signs of ES or EΓ. The retrospective t=-1 hr pre-patterns are a secondary concern about the formation claim; if the stress reconstruction is wrong, the motion-control claim fails even before considering formation. The proposed synthetic-data test would settle whether active stress errors can flip the reconstructed power signs. Until that check is done, the paper should remain conditional, so I do not change the reader's verdict.","tokens_in":20443,"tokens_out":6834,"duration_ms":77754,"concrete_test":"Run a synthetic-data validation with the authors' published monolayer-stress-microscopy code: prescribe a +1/2-defect velocity field and a known active contractile stress field satisfying force balance with traction, input the synthetic traction and velocity into the code, and compare reconstructed PS and EΓ inside the 100×100 µm ROI to ground truth. Active-stress magnitudes should be scanned over the range consistent with the traction/stress ratio in Supplemental Fig. S1. If the reconstructed sign of PS differs from ground truth for any such magnitude, the stress-dominated/traction-dominated classification is not robust. As a cheaper auxiliary check, rerun the reconstruction on the real data with the compatibility constant varied over, say, 0.1–5 and test whether the sign of the ROI-averaged PS is stable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Monolayer stress microscopy (Methods, 'Quantification of Tractions and Stresses') reconstructs the in-plane stress tensor from measured tractions using force balance plus a compatibility equation that, in the authors' words, 'assumes a linear, passive relationship between stress and strain rate,' with the shear-to-bulk ratio fixed at 0.54. The central dichotomy — head-to-tail defects are stress-dominated because PS > 0 and EΓ > 0, while tail-to-head defects are traction-dominated — depends on the sign of the reconstructed stress power and boundary flux. That sign is not protected by the Clapeyron balance ES + ET = EΓ: given any reconstructed σ satisfying ∇·σ = -t/h, that balance is an identity, so the quoted 'couple percent' agreement does not validate the physical interpretation. The cited Zimmermann error analysis (traction at least one-third of stress) bounds global reconstruction error, not local active stresses near defects, where strain-rate gradients and contractile activity are largest. If active stresses are not slaved to strain rate, the reconstructed PS can have the wrong sign, which would invert the energy-source classification for the stress-dominated half of the claim. The traction/velocity angle data are direct and support the traction-propulsion half; the stress-dominated half is the vulnerable piece.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies +1/2 topological defects in confluent MDCK epithelial islands, reporting that defects move in both tail-to-head and head-to-tail directions within the same monolayer. From image correlation velocity fields, traction force microscopy, and monolayer stress microscopy, the authors compute strain rates, stresses, traction-velocity angles, and stress- and traction-based power densities. They find that tail-to-head defects move against the stress gradient with traction aligned to velocity, while head-to-tail defects move with the stress gradient and with traction opposed to velocity. Using a Clapeyron-type power balance over a region of interest, they classify tail-to-head defects as traction-dominated (energy injected via tractions, flowing outward) and head-to-tail defects as stress-dominated (energy flowing inward via stresses and dissipated by tractions). They further report that these strain-rate, stress, traction, and power patterns are already present one hour before defect formation, and that actin stress-fiber orientation differs between the two classes. The central claim is that defect motion direction is controlled by whether energy is injected by tractions or by intercellular stresses, and that coordinated force patterns cause defect formation.","tokens_in":20603,"tokens_out":4610,"duration_ms":51561,"significance":"If the conclusions hold, the paper would provide a notable advance: direct evidence that cell-substrate tractions can act as a propulsive, energy-injecting mechanism rather than a purely passive friction in active nematic models, and a proposal that defect formation is caused by pre-existing coordinated force and motion patterns. The coexistence of both defect-motion directions in the same monolayer is an interesting experimental observation. The paper also ships useful quantitative tools: velocity, traction, and stress fields are computed with standard, openly available code, and the power-balance derivation in Supplemental Note 1 is clear. The main limitations are that the stress field is reconstructed under a passive linear constitutive assumption, and that the pre-formation patterns rely on retrospective alignment to defects identified later. These issues bear directly on the stress-dominated half of the classification and on the formation claim, so the significance is real but conditional.","major_comments":[{"comment":"The stress reconstruction is load-bearing for the claim that head-to-tail defects are stress-dominated, because the signs of PS and EΓ are computed from a reconstructed stress field obtained from a compatibility equation that 'assumes a linear, passive relationship between stress and strain rate' with a chosen shear-to-bulk ratio of 0.54. The Clapeyron balance ES + ET = EΓ is an identity for any symmetric stress field satisfying equilibrium ∇·σ = −t/h; therefore, the reported closure to within a couple percent does not validate the physical interpretation of the sign of EΓ. The cited Zimmermann error analysis bounds global reconstruction errors, not local active-stress errors near defects, where strain-rate gradients and contractile activity are largest. I request robustness tests: (i) repeat the power-density analysis for a range of the ratio around 0.54, (ii) compare the stress-power sign with traction-only measures or with a reconstruction that includes an explicit active stress term, and (iii) apply the Zimmermann-type error model to the actual defect-tail regions with local traction and stress magnitudes. Without such tests, the 'stress-dominated' half of the central dichotomy is not independently established.","section":"Methods, Quantification of Tractions and Stresses; Results, Energy Injection and Dissipation (Fig. 4)"},{"comment":"The claim that strain rate, stress, traction, and power patterns exist at t = −1 hr before defect formation is based on aligning fields to the eventual defect axis and to the eventual direction of motion. This retrospective alignment can create apparent pre-patterns even from random or weakly structured fields if the selection criterion is correlated with the future defect orientation or motion direction. The manuscript does not report controls such as averaging fields around random positions with the same selection criteria, or around defects whose orientation axes are randomly rotated. I ask for explicit null controls and, if possible, a forward-in-time analysis that does not use information about the defect that forms later. This is necessary to support the conclusion that the patterns cause defect formation rather than merely accompany it.","section":"Results, Strain Rates and Stresses near Topological Defects; Methods, Defect Identification"},{"comment":"The classification of defects as 'traction-dominated' versus 'stress-dominated' is partly circular: PT = −t·v/h has a sign fixed by whether traction is aligned or anti-aligned with velocity, which is exactly the criterion used to separate tail-to-head from head-to-tail defects, and PS = σ:ε has a sign fixed by the strain-rate sign that already distinguishes the two classes in Fig. 2. The paper should state explicitly which conclusions are direct kinematic observations (traction-velocity alignment and strain-rate signs) and which require the reconstructed stress. As written, the energy-injection narrative risks restating the classification rather than providing an independent mechanistic explanation. I recommend quantifying how well PS and PT magnitudes and the balance terms predict defect speed or direction beyond the sign of the input fields, and using the actin-stress-fiber data as a more independent test of the proposed mechanism.","section":"Results, Energy Injection and Dissipation (definitions of PS and PT)"}],"minor_comments":[{"comment":"The caption states that <vx> varies from 0 to 20 µm/min, but the main text reports speeds of a few µm/hr; this is likely a units typo and should be corrected to µm/hr.","section":"Supplemental Fig. S6"},{"comment":"The fraction of head-to-tail and tail-to-head defects is shown per treatment, but the number of defects per treatment is not stated; please add n values for the Control, CN02, and CN03 groups so the reader can assess the balance across treatments.","section":"Fig. 1f and main text"},{"comment":"The caption of panel b lists both ES/V and ET/V but the plotted curve appears to be (ES+ET)/V; the caption should be clarified and the individual contributions shown if they are intended.","section":"Supplemental Note 2 and Fig. S11"},{"comment":"The verification criterion for keeping a +1/2 defect is described as 'two counter-rotating vortices on either side of the tail'; please provide quantitative thresholds (e.g., vorticity magnitude and sign at specific positions) so the selection rule is reproducible.","section":"Methods, Quantification of Cell Velocity and Orientation"},{"comment":"The term 'Clapeyron's theorem' is unusual for a power balance of this kind and may be confused with the thermodynamic Clapeyron relation; consider calling it a 'power balance' or 'rate-of-work balance' for clarity.","section":"Supplemental Note 1"}],"recommendation":"major_revision","confidential_remarks":"The paper is ambitious and the experimental dataset is substantial, but the central dichotomy rests on the sign of reconstructed stresses whose constitutive origin is passive and linear. The power-balance identity means that closure alone cannot settle the issue. The pre-formation claim is also vulnerable to selection bias from retrospective alignment. These are fixable with additional controls and sensitivity analyses, so I recommend major revision rather than rejection. If the authors can show that the signs of PS and EΓ are robust to the constitutive assumption and that the t = −1 hr patterns survive null controls, the paper would make a strong contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the paper reports something genuinely new: in the same MDCK island, +1/2 defects move both tail-to-head and head-to-tail, with the two populations nearly balanced (58 vs 53). That directly contrasts with the prior MDCK experiments (Saw 2017, Balasubramaniam 2021), which saw tail-to-head motion as the norm. Second, the traction-velocity angle data are direct and convincing: for tail-to-head defects, traction aligns with velocity along the tail, which is real evidence that substrate traction can act as a propulsive input, not just friction.\n\nThe paper does a lot well. The methods are standard and documented, and they put partial data on figshare and code on GitHub. The power balance ES + ET = EΓ closes to a couple percent, and they also include a full-island noise floor check where the imbalance is about 100 times smaller than the signals of interest. The t = -1 hr retrospective averaging is the one place I would squint: aligning to the future defect axis could create apparent pre-patterns. But they show single-defect examples at t = -1 hr, which mitigates that. The stress-fiber orientation difference between the two classes is a nice independent readout.\n\nThe real soft spot is the stress reconstruction. The stress tensor comes from monolayer stress microscopy, which uses equilibrium plus a compatibility equation that assumes a linear, passive stress-strain-rate relation with the shear-to-bulk ratio fixed at 0.54. The classification of head-to-tail defects as 'stress-dominated' depends on the sign of PS and EΓ, and that sign could be an artifact if active stresses are not slaved to strain rate near the defect. The Clapeyron balance is an identity for any reconstructed σ satisfying equilibrium, so the couple-percent closure is a discretization check, not a physical validation. The Zimmermann error bound they cite is global, not local. I don't think this invalidates the paper, but a referee should ask for a sensitivity analysis on the 0.54 constant and some discussion of active stress magnitude.\n\nAlso, the title says 'control formation,' but the evidence is correlational: patterns precede formation, but nothing is manipulated to show causation. That is a wording overreach, not a data problem.\n\nNet: this is a serious, well-executed experimental contribution. The traction-propulsion half is solid; the stress-dominated half needs extra scrutiny, but the authors already flag the assumption in the methods. Send it to referees, and ask them to focus on the stress reconstruction. The right audience is the active-matter and cell-mechanics community, and they will cite the concurrent-motion observation.","headline":"A novel observation of coexisting defect motion directions with a clear traction-propulsion mechanism, but the stress-based half of the story leans on a constitutive assumption.","tokens_in":21250,"tokens_out":4967,"would_cite":true,"duration_ms":48157,"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":"In MDCK monolayers, +1/2 defects move either tail-to-head (traction-driven) or head-to-tail (stress-driven), and the deciding force patterns exist before the defect forms.","keywords":["topological defects","epithelial monolayers","collective cell migration","traction force microscopy","monolayer stress microscopy","energy injection and dissipation","active nematic","defect formation"],"falsifier":"Measure the stress gradient along the defect tail with a probe that does not assume passive rheology—for example, FRET-based molecular tension sensors in cell-cell adhesions or direct measurement of junction forces—and check whether the gradient remains positive for both head-to-tail and tail-to-head defects. Alternatively, in time-lapse data, ask whether the sign of the average power density in a $100\\times100$ $\\mu$m region one hour before formation predicts each defect's subsequent direction; if it predicts no better than chance, the pre-pattern claim fails.","tokens_in":20145,"feed_emoji":"🧫","tokens_out":5080,"duration_ms":49743,"temperature":0.7,"pith_summary":"Epithelial cells in a confluent monolayer contain comet-shaped flaws in their alignment, called +1/2 topological defects. This paper reports that these defects move in both directions—tail-to-head and head-to-tail—within the same island, and that the direction is set by which interface supplies the energy of motion. When cell-substrate tractions align with cell velocity, they act as a propulsive source and drive defects tail-to-head against the stress gradient; when tractions act as friction, intercellular stresses transmitted by neighboring cells drive defects head-to-tail. The same spatial patterns of strain rate, stress, traction, and power density are present an hour before a defect forms, implying that coordinated force fields cause defect formation rather than merely responding to it. If true, this refocuses theories of collective cell migration on force patterns as the origin of nematic order.","feed_headline":"Tractions can propel, not just resist, epithelial defects","feed_subtitle":"Same monolayer hosts both defect motions; the energy source—traction or stress—sets the direction.","key_machinery":"The analysis rests on a bookkeeping identity for energy flow, Clapeyron's theorem in the form $\\int_\\Omega P_S\\,dA + \\int_\\Omega P_T\\,dA = \\oint_\\Gamma \\sigma \\hat{\\mathbf{n}}\\cdot\\mathbf{v}\\,ds$, with $P_S = \\sigma : \\dot{\\varepsilon}$ the stress power density and $P_T = -\\mathbf{t}\\cdot\\mathbf{v}/h$ the traction power density. It converts separate measurements of velocity, strain rate, traction, and reconstructed stress into signed local rates of energy injection and dissipation, letting each defect be classified as traction-dominated or stress-dominated. The companion structural readout is the angle $\\beta$ between actin stress fibers and the defect tail, which differs by about $90^\\circ$ between the two motion classes.","core_discovery":"The central claim is that +1/2 defects have no unique direction of motion; the balance of energy injection at the cell-substrate interface versus transmission through cell-cell stresses decides. Using velocity fields, traction force microscopy, monolayer stress microscopy, and the power densities $P_S = \\sigma : \\dot{\\varepsilon}$ and $P_T = -\\mathbf{t}\\cdot\\mathbf{v}/h$, the authors find that head-to-tail defects dissipate energy along the tail and are pulled forward by stress transmitted from neighboring cells, whereas tail-to-head defects are propelled by substrate tractions and move against the local stress gradient. They further show that the traction-stress-power patterns pre-exist defect formation by at least one hour, and that actin stress fibers align along the tail only for head-to-tail defects, matching the energy-flow picture. The paper therefore claims that traction is an active, propulsive input and that coordinated patterns of force and motion, rather than preexisting nematic order, are what create +1/2 defects.","pith_inferences":["A direct predictive experiment would be to measure local power densities continuously and register, before any defect appears, whether the sign of the pre-pattern at each future defect location forecasts its eventual direction; the $t=-1$ hr averages imply such forecasts should succeed.","The same energy-balance argument could be applied to $-1/2$ defects and to defects at island boundaries, where the three-fold symmetry or boundary constraints may alter the traction-stress competition.","If traction is an active source near defects, changing substrate stiffness or adhesion ligand density should bias the fraction of tail-to-head defects, a testable mechanical handle not explored here.","The claim that force patterns cause nematic order inverts the usual active-nematic causal arrow; a minimal test would be to ask whether monolayers with randomized initial orientation still nucleate defects when subjected to artificially imposed coordinated traction patterns."],"forward_implications":["A +1/2 defect's direction of motion cannot be used as a readout of whether the monolayer is extensile or contractile; the same stress state accompanies both directions.","Cell-substrate traction must be treated as an active, energy-injecting field in models of epithelial monolayers, not merely as passive friction.","Patterns of force, strain rate, and energy injection present before defect formation imply that defects arise from coordinated cell motion rather than from a preexisting tendency for nematic alignment.","Energy flows along defect tails, so defects act as local conduits that either receive power from neighbors (head-to-tail) or send power out to neighbors (tail-to-head).","Actin stress fiber orientation at the defect tail distinguishes the two regimes and may serve as a structural marker of which energy source dominates."],"supporting_citations":[{"why":"Supplies the prior observation that +1/2 defects in epithelia move tail-to-head and sets the comparison baseline for stress and strain-rate patterns near defects.","marker":"[2]"},{"why":"Reports the puzzling result that contractile cells can form extensile-style monolayers and that tail-to-head defects move against stress gradients, which this paper reinterprets.","marker":"[10]"},{"why":"Provides the low-density island experimental system on 6 kPa substrates and shows that tractions and stress fibers control cell shape and motion.","marker":"[21]"},{"why":"Introduces monolayer stress microscopy, the method used here to reconstruct intercellular stress from traction data.","marker":"[34]"},{"why":"Documents the assumptions and accuracy of monolayer stress microscopy, including the dimensionless constant set to 0.54.","marker":"[35]"},{"why":"Supplies the error analysis showing that reconstructed stresses are reliable when traction magnitude is at least one third of stress magnitude.","marker":"[45]"},{"why":"Offers a theoretical alternative in which cell-shape fluctuations generate tail-to-head defects in contractile systems, against which the traction-propulsion picture is compared.","marker":"[41]"},{"why":"Proposes two distinct orientation fields for cell shape and actomyosin stress fibers, used here to interpret the observed stress-fiber alignment differences.","marker":"[42]"},{"why":"Shows that cell-body orientation need not align with first principal stress, supporting the claim that defect motion is not a reliable indicator of the stress state.","marker":"[33]"},{"why":"Shows how fluctuating polar forces such as tractions can produce local nematic order and extensile-like behavior, supporting the idea that traction is an active source.","marker":"[22]"}],"fun_headline_variants":["Defect motion direction set by traction vs stress energy","Tractions steer epithelial defect motion both ways","Energy source decides which way +1/2 defects move","Force patterns predict defect formation and movement"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The computed intercellular stress fields are trusted, even though they are reconstructed from traction data by assuming a linear, passive relation between stress and strain rate; if real active stresses differ near defects, the stress-gradient and stress-power classification could be wrong.","fun_headline_variants_meta":{"raw":{"variants":["Defect motion direction set by traction vs stress energy","Tractions steer epithelial defect motion both ways","Energy source decides which way +1/2 defects move","Force patterns predict defect formation and movement"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000233,"raw_usage":{"total_tokens":1500,"prompt_tokens":957,"completion_tokens":543,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":573,"completion_tokens_details":{"reasoning_tokens":484}},"tokens_in":573,"tokens_out":543,"duration_ms":5197,"temperature":1.0,"reasoning_tokens":484,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:25:14.042238+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the stress gradient along the defect tail with a probe that does not assume passive rheology—for example, FRET-based molecular tension sensors in cell-cell adhesions or direct measurement of junction forces—and check whether the gradient remains positive for both head-to-tail and tail-to-head defects. Alternatively, in time-lapse data, ask whether the sign of the average power density in a $100\\times100$ $\\mu$m region one hour before formation predicts each defect's subsequent direction; if it predicts no better than chance, the pre-pattern claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the prior observation that +1/2 defects in epithelia move tail-to-head and sets the comparison baseline for stress and strain-rate patterns near defects."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the puzzling result that contractile cells can form extensile-style monolayers and that tail-to-head defects move against stress gradients, which this paper reinterprets."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the low-density island experimental system on 6 kPa substrates and shows that tractions and stress fibers control cell shape and motion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces monolayer stress microscopy, the method used here to reconstruct intercellular stress from traction data."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the assumptions and accuracy of monolayer stress microscopy, including the dimensionless constant set to 0.54."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the error analysis showing that reconstructed stresses are reliable when traction magnitude is at least one third of stress magnitude."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Offers a theoretical alternative in which cell-shape fluctuations generate tail-to-head defects in contractile systems, against which the traction-propulsion picture is compared."},{"cited_title":"Cellular dynamics emerging from turbulent flows steered by active filaments","cited_arxiv_id":"2501.05971","evidence_quote":"Proposes two distinct orientation fields for cell shape and actomyosin stress fibers, used here to interpret the observed stress-fiber alignment differences."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that cell-body orientation need not align with first principal stress, supporting the claim that defect motion is not a reliable indicator of the stress state."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how fluctuating polar forces such as tractions can produce local nematic order and extensile-like behavior, supporting the idea that traction is an active source."}],"review_version":1}