Pith. sign in

REVIEW 3 major objections 4 minor 30 references

Collective cell migration of epithelial cells driven by chiral torque generation

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper proposes that a spatial gradient of cellular chiral torque makes epithelial cells migrate unidirectionally, with the migration speed determined solely by the average cell shape index in the boundary layers.

desk verdict A genuinely new chiral vertex model with a validated mean-field speed, but the 'scaled solely by shape index' claim is narrower than the data. read the letter →

arxiv 1909.02771 v1 pith:CWJYJQZL submitted 2019-09-06 cond-mat.soft physics.bio-ph

Takaki Yamamoto , Tetsuya Hiraiwa , Tatsuo Shibata This is my paper · ORCID
classification cond-mat.softphysics.bio-ph
keywords chiraltorquecollectivecellmigrationvertexmodelshapeindexT1transitionliquid-solidleft-rightasymmetryepithelialtissue
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to close a gap: experiments show that individual cells can generate chiral torques and that tissues display chiral morphogenesis, but no theory connected the two. It adds a chiral torque to each cell in a standard vertex model of a confluent epithelial sheet. When the torque strength is uniform, the model produces bidirectional flows driven at the boundaries; when the torque strength varies linearly across the sheet, cells migrate unidirectionally, perpendicular to the gradient, at a speed set by the mean-field relation $v = 3\lambda l^2/2\eta$. The central claim is that, despite the many mechanical parameters, the bulk migration speed collapses onto a single curve when plotted against the average cell shape index in the boundary layers, the same index that marks the liquid-to-solid transition in confluent tissues. A sympathetic reader would care because it offers a concrete, testable bridge from single-cell chirality to tissue-level left-right asymmetry.

What carries the argument

The machinery is the two-dimensional cell vertex model, in which cells are polygons and vertices move under force balance between friction, a potential energy from area and perimeter elasticity, and a chiral torque force $\mathbf{T}_i = \sum_\alpha \nu_\alpha (\mathbf{r}_i - \mathbf{r}_g^\alpha) \times \mathbf{n}$ exerted around each cell's area centroid. T1 topological transitions flip short edges to allow neighbour exchanges, and they are activated by a coloured, time-correlated line-tension fluctuation. The argument is carried by a mean-field three-cell calculation for a regular hexagonal vertex, which yields the net force $(3\lambda l^2/2, 0)$ under a linear torque gradient, and by the cell shape index $q = P/\sqrt{A}$, which the paper uses as the single order parameter that collapses the velocity data. The shape index also names the liquid-to-solid transition in confluent tissues, linking the migration mechanism to rigidity.

What would settle it

Run the model at several torque-gradient slopes $\lambda$, line-tension correlation times $\tau$, and tissue widths $L_y$, and check whether all bulk-velocity curves still collapse onto a single curve when plotted against $\langle q\rangle_{BL}$; lack of collapse would refute the claim that the shape index alone sets the migration speed.

Watch

Extended reading notes

Core claim

The paper's central claim is that a gradient in the strength of cellular chiral torque makes a confluent epithelial sheet migrate unidirectionally, perpendicular to the gradient, and that the resulting bulk migration speed is fixed by a single structural quantity: the average cell shape index in the boundary layers. In a mean-field three-cell calculation, the chiral torque forces combine to give a net force $(3\lambda l^2/2, 0)$ on a vertex, predicting $v_{\mathrm{theory}} = 3\lambda l^2/(2\eta)$ for hexagonally packed cells, and the simulations approach this value as system size grows. With line-tension fluctuations present, velocity data for different noise strengths $\sigma$ and target shape indices $P_0$ collapse onto one curve when plotted against $\langle q\rangle_{BL}$, which the paper interprets as the shape index encoding both the bulk torque force and the height of the T1 rearrangement barrier. The paper also establishes two supporting facts: a homogeneous torque produces only bidirectional, boundary-localized flows, and without line-tension fluctuations cells deform but never flow.

Load-bearing premise

The universal speed scaling rests on the assumption that the averaged boundary-layer shape index carries all the information about both the driving torque and the T1 rearrangement barrier, so the bulk migration speed is a single-valued function of that index alone.

Editorial extensions

If this is right

  • Uniform chiral torque generation produces only bidirectional, boundary-localized flows in a sufficiently wide tissue, so bulk unidirectional migration requires a torque gradient.
  • Under a torque gradient, the migration direction is perpendicular to the gradient, and the mean-field speed grows linearly with the slope $\lambda$.
  • Continuous migration requires T1 rearrangements, which the line-tension fluctuations provide; without them, the tissue only deforms.
  • If the shape-index collapse is robust, measuring the average boundary-layer shape index of an epithelial sheet could predict its collective migration speed without knowing the detailed noise or target-perimeter parameters.
  • The mechanism offers a route from cell-autonomous chirality to left-right asymmetric tissue behaviours such as tube twisting and unidirectional epithelial flow during development.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the collapse by $\langle q\rangle_{BL}$ holds beyond the tested parameter range, then the shape index acts as a practical readout: experimentalists could infer expected migration speed from static shape measurements alone.
  • The scaling argument connects naturally to the rigidity transition of confluent tissues, suggesting that the same order parameter might control both solid-fluid behaviour and chiral migratory response.
  • Tissue width should determine which regime is observed: narrow sheets may show boundary-dominated bidirectional flows even under a gradient, while wide sheets allow bulk unidirectional migration once boundary effects are suppressed.
  • A testable extension would be to impose a biochemical gradient of a regulator of actomyosin chirality in an epithelial monolayer and ask whether cells migrate perpendicular to the gradient at a speed correlated with the local shape index.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes an extension of the two-dimensional cell vertex model in which each cell generates a chiral torque, modeled by a force on each vertex proportional to the lever arm from the cell centroid (Eq. 2). For a spatially homogeneous torque strength, simulations show bidirectional boundary-localized flows that decay in the bulk. When the torque strength varies linearly along the y-direction, ν_α = −λ(y_g^α − L_y), the tissue migrates unidirectionally along x, perpendicular to the torque gradient. The authors derive a mean-field speed v_theory = 3λl^2/2η for a regular hexagonal configuration and show that measured speeds approach this value as the system size increases (Fig. 4(b)). They further report that the bulk velocity normalized by v_theory collapses onto a single curve when plotted against the cell shape index ⟨q⟩_BL averaged in boundary layers, for different values of the line-tension noise σ and the target shape index P0 (Fig. 5). The paper interprets this as evidence that the migration speed is 'scaled solely' by a structural order parameter for the liquid-to-solid transition.

Significance. If the results hold, the paper provides a minimal and appealing physical mechanism connecting single-cell chirality to tissue-scale left-right asymmetry and directional collective migration, a topic of broad interest in developmental biology and active matter. The strengths of the paper are its transparent model, the clean mean-field prediction that is numerically verified, the use of no fitted free parameters in the comparison, and the falsifiable prediction that a torque gradient drives perpendicular unidirectional flow. The mean-field derivation, the numerical validation for large systems, and the demonstration that T1 rearrangements (enabled by line-tension fluctuations) are essential for continuous flow are solid contributions. The weaker part is the universal scaling claim, which is broader than the evidence presented.

major comments (3)
  1. [Mean-field model (near Fig. 4(a))] The result f = (3λl^2/2, 0) is introduced with 'After a straightforward calculation,' but the calculation is not shown in the main text or the Supplemental Material. Because v_theory = 3λl^2/2η is the reference speed used to normalize all measured velocities (Fig. 4(b) and Fig. 5), this derivation should be presented explicitly, at least in the Supplemental Material, so that the reader can verify the assumptions about the three-cell geometry and the treatment of the torque gradient.
  2. [Fig. 5 and abstract (scaling claim)] The claim that the migration speed is 'scaled solely by' the cell shape index is supported only by a data collapse for variations of σ and P0 at fixed λ = 0.01, Ny = 40, K = 10, and τ = 1. Because v_theory itself is proportional to λ, normalizing by v_theory does not remove a possible separate dependence of the migration speed on the torque-gradient strength. Additional simulations varying λ (and ideally K, τ, and system size) are needed before the 'solely' statement in the abstract can be claimed; alternatively, the claim should be qualified to the tested parameter ranges.
  3. [Mechanistic explanation following Fig. 5] The explanation that 'the torque force depends on the cell shape' and is therefore determined by the scalar quantity ⟨q⟩_BL is not derived from Eq. (2). The torque force on a vertex is a vector sum of terms ν_α (r_i − r_g^α) × n, which depend on the precise centroid and vertex positions, not only on the dimensionless shape index q. No argument is given that configurations with the same q produce the same net torque force. Without such an argument or a test that varies the torque gradient, the uniqueness of ⟨q⟩_BL as the sole determining variable of the bulk velocity is not established.
minor comments (4)
  1. [Title (arXiv header)] The title in the arXiv header contains a typo, 'chi ral torque generation'; the running title and metadata should be corrected.
  2. [Fig. 5 caption] The caption should clarify that the inset reports Vave (averaged over the entire sheet) while the main panel reports Vbulk (averaged over bulk layers only), and should define which quantity is used in the collapse so the reader can interpret the two panels consistently.
  3. [Supplemental Material, §IV (boundary layers)] The boundary layers are defined operationally from the velocity profile using a 3σ criterion, and the top and bottom layers are excluded from ⟨q⟩_BL because they are outliers. A brief discussion of how sensitive the collapse in Fig. 5 is to this operational definition would help rule out artifacts of the layer-selection procedure.
  4. [Throughout] The noise correlation time is fixed at τ = 1 without a stated justification. Since τ affects the T1 rearrangement rate, a sentence acknowledging that the scaling claim has not been tested in τ would be appropriate, especially given the abstract's universal phrasing.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the mean-field velocity prediction is derived from the model and independently validated by simulation, and the scaling collapse is an emergent observation rather than a fitted construction.

full rationale

The paper's central quantitative prediction, v_theory = 3λl²/2η, is obtained by applying the torque-force definition in Eq. (2) to a regular hexagonal three-cell vertex configuration, and it is then tested numerically by switching off line-tension fluctuations, suppressing boundary torque forces, and varying system size and λ (Fig. 4). The numerics approach the theoretical value as Ny increases, and the residual deviation is attributed to cellular deformation, which the paper independently verifies through the shape-index measurement. No parameter is fitted to force agreement. The universal scaling collapse in Fig. 5 is also not circular: ⟨q⟩_BL is measured from the same simulations, and the collapse is presented as an empirical observation with a qualitative mechanistic explanation, not as a quantity defined to reproduce the velocity. The paper does invoke a symmetry argument from prior work by overlapping authors ([12]) to motivate breaking y-symmetry, but the unidirectional migration result does not reduce to that citation: it is demonstrated by direct numerical simulation and by the explicit force-balance calculation. Concerns that the collapse is only shown for λ = 0.01 or that the bulk torque force is asserted to be a function of the scalar shape index are legitimate questions about the scope and robustness of the scaling claim, but they are not circularity. There is no self-definitional step, no fitted input renamed as prediction, and no load-bearing self-citation chain.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The central claims rest on a base vertex model and several modeling choices: the specific torque force form, the linear torque gradient, the fluctuation model, and the T1 transition rule. None of these are derived from a more microscopic actomyosin theory, so they are axioms of the model. The paper does not fit any parameters to data.

free parameters (8)
  • Area elasticity K = 10
    Set by hand to prevent large shape deformation; not varied in the main scaling study.
  • Target shape index P0 = 3.54 (default), 2.90 (mean-field test), varied in Fig. 5 inset
    Controls cell shape and T1 barrier; the key mechanical parameter varied to test the scaling.
  • Line tension noise strength σ = 0.1-0.3
    Controls fluctuation-driven cell rearrangements; varied to test the scaling collapse.
  • Noise correlation time τ = 1
    Set to 1 in this letter, simplifying the colored noise.
  • Friction coefficient η = 1 (bulk), 10000 (boundaries in gradient runs)
    Sets the drag; the mean-field speed v_theory is inversely proportional to η.
  • Torque gradient λ = 0.01 (scaling study), varied 0.01-0.2 (Fig. 4b)
    The driving force in the model; normalized velocity V_bulk/v_theory is plotted against the shape index to test universality.
  • T1 threshold l_th = 0.03
    Edge length below which a T1 flip is performed; a numerical parameter.
  • Chiral torque strength ν (homogeneous case) = 0.2 or 1.0
    Used for the homogeneous torque case to demonstrate bidirectional boundary flow.
assumptions (6)
  • domain assumption The cell vertex model energy (area elasticity, perimeter elasticity, line tension) describes confluent epithelial tissue mechanics.
    Used as the base model (Eq. 1), cited to Nagai and Honda (2001). None of these energy terms are derived from a more fundamental theory.
  • ad hoc to paper The chiral torque generated by a cell can be represented as a force on each vertex proportional to the lever arm from the cell centroid, T_i = sum ν_α (r_i - r_g^α) × n.
    Introduced in Eq. (2) with no derivation from actomyosin mechanics; the form is chosen as the simplest torque force compatible with the vertex model.
  • domain assumption Line tension fluctuates as colored Gaussian noise with correlation time τ = 1.
    Used to drive T1 rearrangements; supported by experimental report in ref [21] but simplified to τ = 1.
  • domain assumption Cell rearrangements occur through T1 transitions when an edge shrinks below l_th = 0.03.
    Standard vertex model rule; the threshold is arbitrary but small enough not to affect the physics.
  • ad hoc to paper The torque strength varies linearly as ν = -λ(y_g - Ly), a simple gradient that breaks y-symmetry.
    The simplest linear form; the biological origin is proposed to be a gradient of regulatory molecules, but no specific mechanism is modeled.
  • domain assumption The cell shape index q = P/√A controls the energy barrier for T1 transitions and the liquid-to-solid transition in confluent tissues.
    Taken from prior work by Bi et al. (refs 23, 25, 26). This is central to the interpretation of the scaling collapse.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Collective cell migration of epithelial cells driven by chiral torque generation." pith.science (2026). https://pith.science/paper/CWJYJQZL

@misc{pith2026190902771,
  author       = {Pith},
  title        = {Pith review of: Collective cell migration of epithelial cells driven by chiral torque generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CWJYJQZL}},
  note         = {Machine review of arXiv:1909.02771}
}
read the original abstract

Various multicellular tissues show chiral morphology. Experimental studies have shown this can originate from cell chirality. However, no theory has been proposed to connect the cellular chiral torque and multicellular chiral morphogenesis. We propose a model of confluent tissue dynamics with cellular chiral torque. We found that cells migrate unidirectionally under a gradient of cellular chiral torque. While the migration speed varies depending on the tissue's mechanical parameters, it is scaled solely by a structural order parameter for liquid-to-solid transition in confluent tissues.

Figures

Figures reproduced from arXiv: 1909.02771 by the authors.

Figure 1
Figure 1. FIG. 1. (a) Schematic of the torque forces exerted on the [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. FIG. 3. (a) Time evolution of a cellular configuration in a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. FIG. 4. (a) A cellular configuration for the theoretical anal [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: ). Here, we set Ny = 40, λ = 0.01 and applied the same boundary conditions that were used in [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 28 canonical work pages

  1. [1]

    Mogilner and B

    A. Mogilner and B. Fogelson, Curr. Biol. 25, R501 (2015)

    work page 2015

  2. [2]

    Inaki, J

    M. Inaki, J. Liu, and K. Matsuno, Philos. Trans. Royal Soc. B 371, 20150403 (2016)

    work page 2016

  3. [3]

    S. R. Naganathan, T. C. Middelkoop, S. F¨ urthauer, and S. W. Grill, Curr. Opin. Cell Biol. 38, 24 (2016)

    work page 2016

  4. [4]

    L. Q. Wan, A. S. Chin, K. E. Worley, and P. Ray, Phil. Trans. R. Soc. B 371, 20150413 (2016)

    work page 2016

  5. [5]

    Hirokawa, Y

    N. Hirokawa, Y. Tanaka, and Y. Okada, Cold Spring Harb. Perspect. Biol. 1, a000802 (2009)

    work page 2009

  6. [6]

    Lebreton, C

    G. Lebreton, C. G´ eminard, F. Lapraz, S. Pyrpassopou- los, D. Cerezo, P. Sp´ eder, E. M. Ostap, and S. Noselli, Science 362, 949 (2018)

    work page 2018

  7. [7]

    Taniguchi, R

    K. Taniguchi, R. Maeda, T. Ando, T. Okumura, N. Nakazawa, R. Hatori, M. Nakamura, S. Hozumi, H. Fujiwara, and K. Matsuno, Science 333, 339 (2011)

    work page 2011

  8. [8]

    Suzanne, A

    M. Suzanne, A. G. Petzoldt, P. Sp´ eder, J.-B. Coutelis, H. Steller, and S. Noselli, Curr. Biol 20, 1773 (2010)

    work page 2010

Show all 30 references
  1. [9]

    K. Sato, T. Hiraiwa, E. Maekawa, A. Isomura, T. Shi- bata, and E. Kuranaga, Nat. Commun. 6, 10074 (2015)

    2015

  2. [10]

    Guillot and T

    C. Guillot and T. Lecuit, Science 340, 1185 (2013)

    2013

  3. [11]

    Hiraiwa, E

    T. Hiraiwa, E. Kuranaga, and T. Shibata, Front. Cell Dev. Biol. 5, 66 (2017)

    2017

  4. [12]

    K. Sato, T. Hiraiwa, and T. Shibata, Phys. Rev. Lett. 115, 188102 (2015)

    2015

  5. [13]

    Tamada, S

    A. Tamada, S. Kawase, F. Murakami, and H. Kamiguchi, J. Cell Biol. 188, 429 (2010)

    2010

  6. [14]

    S. R. Naganathan, S. F¨ urthauer, M. Nishikawa, F. J¨ ulicher, and S. W. Grill, eLife 3, e04165 (2014)

    2014

  7. [15]

    Yamanaka and S

    H. Yamanaka and S. Kondo, Genes Cells 20, 29 (2015)

    2015

  8. [16]

    Y. H. Tee, T. Shemesh, V. Thiagarajan, R. F. Hari- adi, K. L. Anderson, C. Page, N. Volkmann, D. Hanein, S. Sivaramakrishnan, M. M. Kozlov, and A. D. Bershad- sky, Nat. Cell Biol. 17, 445 (2015)

    2015

  9. [17]

    F¨ urthauer, M

    S. F¨ urthauer, M. Strempel, S. W. Grill, and F. J¨ uliche r, Eur. Phys. J. E 35, 89 (2012)

    2012

  10. [18]

    F¨ urthauer, M

    S. F¨ urthauer, M. Strempel, S. W. Grill, and F. J¨ uliche r, Phys. Rev. Lett. 110, 048103 (2013)

    2013

  11. [19]

    F¨ urthauer, M

    S. F¨ urthauer, M. Neef, S. W. Grill, K. Kruse, and F. J¨ ulicher, New J. Phys. 14, 023001 (2012)

    2012

  12. [20]

    Nagai and H

    T. Nagai and H. Honda, Philos. Mag. B 81, 699 (2001)

    2001

  13. [21]

    Curran, C

    S. Curran, C. Strandkvist, J. Bathmann, M. de Gennes, A. Kabla, G. Salbreux, and B. Baum, Dev. Cell 43, 1 (2017)

    2017

  14. [22]

    Farhadifar, J.-C

    R. Farhadifar, J.-C. R¨ oper, B. Aigouy, S. Eaton, and F. J¨ ulicher, Curr. Biol.17, 2095 (2007)

    2007

  15. [23]

    D. Bi, J. H. Lopez, J. M. Schwarz, and M. L. Manning, Nat. Phys. 11, 1074 (2015)

    2015

  16. [24]

    See Supplemental Material for movies of cellular migra - tion, definition of the area centroid of a cell, preparation of the initial cellular configuration, cellular deformatio n when σ = 0, definition of the boundary layers, relation- ship between ⟨qα ⟩BL and ⟨qα ⟩BU

  17. [25]

    J.-A. Park, J. H. Kim, D. Bi, J. A. Mitchel, N. T. Qazvini, K. Tantisira, C. Y. Park, M. McGill, S.-H. Kim, B. Gweon, J. Notbohm, R. Steward Jr, S. Burger, S. H. Randell, A. T. Kho, D. T. Tambe, C. Hardin, S. A. Shore, E. Israel, D. A. Weitz, D. J. Tschumperlin, E. P. Henske, ...

    2015

  18. [26]

    D. Bi, X. Yang, M. C. Marchetti, and M. L. Manning, Phys. Rev. X 6, 021011 (2016)

    2016

  19. [27]

    Etienne-Manneville and A

    S. Etienne-Manneville and A. Hall, Nature 420, 629 (2002) . arXiv:1909.02771v1 [cond-mat.soft] 6 Sep 2019 SUPPLEMENTAL MATERIAL Collective cell migration of epithelial cells driven by chi ral torque generation Takaki Yamamoto,1, ∗ Tetsuya Hiraiwa, 2, 3 and Tatsuo Shibata 1, † ...

    2002 arXiv

  20. [28]

    We calculated the approximate bulk velocity V ′ bulk and the approximate standard deviation σ′ bulk of the bulk velocity by averaging the velocity of the 15th ∼ 24th layers

  21. [29]

    We smoothed the velocity profile to obtain v = vsm(i) by a simple moving average, where the mean is taken from two data on either side of a central value

  22. [30]

    Also, Nt is determined as the minimum integer i > 24 which satisfies either vsm(i) < V ′ bulk −3σ′ bulk or vsm(i) > V ′ bulk + 3σ′ bulk

    We determined Nb as the maximum integer i < 15 which satisfies either vsm(i) < V ′ bulk − 3σ′ bulk or vsm(i) > V ′ bulk +3σ′ bulk. Also, Nt is determined as the minimum integer i > 24 which satisfies either vsm(i) < V ′ bulk −3σ′ bulk or vsm(i) > V ′ bulk + 3σ′ bulk. In Fig. S3(...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.