{"id":"b72b2a3a-1112-4ae4-896d-401037226ae6","arxiv_id":"2507.00245","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Off-tangent chiral active forces are sufficient to produce rotating active nematic order in simulations, with handedness tunable by filament stiffness, supported by gliding assays.","lead":"This paper shows that microtubules driven by kinesin motors can form large rotating patterns because the motors push at a slight angle, not because the filaments are curved. The direction and speed of rotation can be reversed by making the filaments stiffer or softer, a prediction tested in experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Taxol handedness interpretation is the load-bearing weak point: the experimental claim of stiffness-tunable handedness relies on an admitted double-sign-flip (α<0 plus low-κ̃) that the authors concede an alternative α-sign scenario would invalidate.","rationale":"The paper's central claim has two parts: a simulation-based sufficiency argument (off-tangent chiral active forces generate time-cholesteric order without shape chirality) and an experimental prediction (handedness and angular speed tunable by filament stiffness). The simulation part is credible and is the main contribution; it does not depend on the contested experimental assignments. The experimental part, however, is only as strong as the taxol interpretation. The authors assign taxol-stabilized MTs α<0 because of the expected 12-protofilament excess, and low κ̃ because taxol lowers bending rigidity; GMPCPP MTs are assigned α>0 and high κ̃. Since both rotate CCW, the taxol case requires two sign reversals to match the model. The paper explicitly admits that if kinesin sidestepping makes the effective α positive for all protofilament numbers, the proposal that taxol rotates opposite to its chiral self-propulsion is invalidated, and that taxol could instead be re-classified as high-κ̃. This flexibility means the experimental data do not pin down the mechanism; they are consistent with the model only under a particular, admittedly uncertain assignment of two parameters. The proximity of the estimated κ̃≈250 to the model threshold κ̃_c≈200 further weakens the rigidity-based placement. A second, related gap is that single GMPCPP MTs in the experiments rotate CCW at 0.029°/s, faster than the measured collective rotation, whereas the model's high-κ̃ single filaments show no significant rotation; this discrepancy is not resolved but is secondary to the taxol sign issue. The reader's weakest_assumption correctly identifies the double-sign-flip as the central fragile premise. My recommendation is to keep the CONDITIONAL verdict: the simulation contribution stands, but the experimental claim of stiffness-tunable handedness should be treated as a suggestive consistency argument pending independent determination of the effective α sign and the κ̃ regime of taxol-stabilized MTs.","tokens_in":25892,"tokens_out":6498,"duration_ms":75991,"concrete_test":"Track isolated taxol-stabilized microtubules gliding in the same supported-lipid assay used for Fig. 4D and measure their mean angular turning rate. Under the paper's assignment (α<0, low-κ̃), the model's single-filament regime (Fig. 2D) predicts CW rotation, opposite to GMPCPP's CCW. If taxol MTs instead rotate CCW (or not at all), the α<0 and/or low-κ̃ assignment fails, falsifying the double-sign-flip explanation for the observed CCW collective rotation. A complementary check: compute the effective α for the actual kinesin-1 construct using published sidestepping bias data for N=12 and N=14 MTs; if α>0 for both, the taxol α<0 premise is wrong.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Simulation evidence that off-tangent chiral active forces suffice for time-cholesteric order is solid. However, the paper's experimental validation of stiffness-tunable handedness rests entirely on interpreting the CCW rotation of taxol-stabilized MTs as α<0 (12-protofilament excess) combined with low bending rigidity (κ̃≲200), while GMPCPP MTs are α>0 and high-κ̃. Both observations are CCW, so the taxol case requires two sign flips. This assignment is not independently anchored: the sign of the effective α is ambiguous because kinesin sidestepping is biased leftward and may produce α>0 for all N; the authors write explicitly 'We cannot rule out this hypothesis. If true, this would invalidate our proposal that taxol-stabilized microtubules rotate opposite to the handedness of their chiral self-propulsion.' They then add that if α>0 for all N, taxol could simply be re-classified as high-κ̃, showing the κ̃ assignment is adjustable rather than measured. The estimated κ̃≈250 for typical MTs sits close to the model's threshold κ̃_c≈200, and no measurement places taxol firmly below threshold. Since the headline prediction (handedness tunable by stiffness) is experimentally supported only by this double-sign-flip, the experimental component of the central claim is not secure; it is a consistency argument, not a confirmation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript combines gliding assays of GMPCPP- and taxol-stabilized microtubules on glass and lipid substrates with Brownian dynamics simulations of semiflexible filaments propelled by a chiral active force tilted at a fixed angle α to the local tangent. The authors report that the emergent long-range nematic state rotates coherently (a 'time cholesteric'), that the rotation rate anti-correlates with nematic order in experiments, that chiral active forces without shape chirality are sufficient to produce such rotating states in simulations, and that the collective angular velocity reverses sign as bending stiffness κ̃ is varied below a threshold κ̃≈200. They present pair-collision simulations and an iterative Markov-chain extrapolation to support a collision-driven mechanism, and they interpret the CCW rotation of both GMPCPP- and taxol-stabilized microtubules on lipid substrates as resulting from a double sign flip in α and κ̃ for taxol.","tokens_in":26207,"tokens_out":10884,"duration_ms":121716,"significance":"The central simulation result—that off-tangent chiral active forces suffice to produce coherently rotating nematic order, with a non-trivial stiffness-controlled handedness reversal—is a credible and valuable contribution to active matter. The pair-collision analysis with an explicit extrapolation to bulk rotation is a strong feature, as is the experimental demonstration that filament-shape curvature cannot account for the observed rotation. The stiffness-tuning prediction appears to have been made before the taxol experiments, which mitigates circularity concerns. However, the experimental evidence for stiffness-tunable handedness is only a consistency argument, not a confirmation, and the taxol interpretation carries an admitted double-sign-flip assignment.","major_comments":[{"comment":"The experimental support for the stiffness-tunable handedness prediction rests on a double-sign-flip assignment that is not independently anchored. The paper assumes taxol-stabilized microtubules have α<0 (12-protofilament excess) and assigns them to the low-κ̃ regime because the literature value κ̃≈250 is 'notably close' to the threshold κ̃≈200; neither the protofilament distribution nor the effective bending rigidity of the taxol microtubules in this assay is measured. Both observed states rotate CCW (Fig. 4D), so the taxol datum is equally consistent with the alternative, explicitly conceded hypothesis that kinesin sidestepping makes α>0 for all protofilament numbers and that taxol microtubules belong to the high-κ̃ regime; under that hypothesis there is no handedness reversal to explain. Since the paper states 'We cannot rule out this hypothesis. If true, this would invalidate our proposal...', the experiments should be described as a consistency check rather than as evidence for the stiffness-tuned handedness, unless direct measurements of the protofilament distribution and effective stiffness are added.","section":"Experimental evidence for collision-driven, rigidity-modulated collective rotation (Fig. 4D,E; Discussion)"},{"comment":"The reported anti-correlation between nematic order parameter and rotation rate is confounded by the simultaneous density change. The authors decreased the microtubule surface density from 0.20±0.02 to 0.13±0.01 filaments/µm² to obtain the lower-order state; lower density generally decreases, not increases, collision frequency, so the statement that 'collisions are expected to be more frequent when nematic order is lower' is not established by this manipulation. The observed increase in rotation rate could reflect a density-dependent change in the active nematic state rather than a collision-driven torque. A test at fixed density, or a direct measurement of collision events, is needed before this experimental result can be used to support the collision-driven mechanism.","section":"Collective rotation emerges from interactions, not filament curvature (Fig. 1E)"},{"comment":"The model's parameter mapping to the experiments is strained by a single-filament discrepancy. The simulations show that high-κ̃ filaments, which are stated to correspond to the experimental microtubules, exhibit no significant average rotation, and the paper uses this to argue that the collective rotation arises from filament interactions. However, the single-filament gliding experiments on GMPCPP microtubules measure a statistically significant CCW turning bias (p=5.75×10^-11, 0.029 deg/s). If the experimental filaments rotate, they are not in the simulated high-κ̃ single-filament regime, so the claim that individual rotation does not contribute to the experimental collective rotation is not supported. Please report the simulated single-filament rotation rate at κ̃=300 in the same units as the experiments, or otherwise reconcile the measured individual turning with the model.","section":"Modeling supports chiral self-propulsion as driver of rotation, via collisions (Fig. 2D vs. Fig. 1F)"}],"minor_comments":[{"comment":"The taxol lipid-bilayer experiments are described inconsistently: the main text reports n=2 experiments with density≈0.45 filaments/µm², while the Fig. 4D caption reports three experiments and mentions a second density of ≈0.95 filaments/µm²; please reconcile these numbers and describe the statistical treatment of multiple fields of view.","section":"Fig. 4D caption vs. main text"},{"comment":"Reference [29] is cited for the claim that nematic ordering correlates with surface density and crowding-agent concentration, but the listed reference (Saito et al., RSC Adv. 2017) is about glyme-based electrolytes and appears unrelated; this citation should be corrected.","section":"References [29]"},{"comment":"The Supplementary Information contains a duplicated phrase: 'When when chirality is introduced' in the discussion of stationary matrices; please fix this typo.","section":"Supporting Information, Section 2"},{"comment":"The sentence 'We find no overlap of the high-κ̃ regime in which LRO states emerge and the low-κ̃ regime in which individual filaments rotate with a consistent handedness' is potentially confusing because 'overlap' is used to describe disjoint intervals in κ̃; consider rewording to 'there is no κ̃ value belonging to both regimes.'","section":"Discussion"}],"recommendation":"major_revision","confidential_remarks":"This is an honest manuscript with a clearly described suite of simulation results and an experimental section whose limitations are acknowledged. The central sufficiency claim is sound and worth publishing. The main issue is that the abstract and section titles currently overstate the experimental confirmation of stiffness-tunable handedness; the admitted taxol caveat should be moved into the Results and the claims softened if no new measurements are added. The paper fits the journal's scope and does not raise novelty concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [colleague],\n\nThe thing to know: the simulation result is solid, and the experimental validation is shakier than the paper would like, in exactly one place—the taxol-stabilized microtubules are used as confirmation of the stiffness-tuned handedness reversal, but the interpretation rests on a double-sign-flip that the authors themselves admit is not secure. Read it for the simulations; treat the taxol story as a consistency argument.\n\nWhat's new: off-tangent chiral active forces, with no shape chirality, are sufficient to produce a coherently rotating active nematic. I believe that is new. The handedness reversal as bending rigidity crosses a threshold is an emergent simulation result, not put in by hand, and the pair-collision Markov scheme is a nice piece of work: it predicts the sign change of the per-collision rotation from binary scattering data, even though the predicted threshold (~κ̃=125) is a factor of two below the bulk simulation threshold (~200), which they acknowledge. The lipid-bilayer experiments on both GMPCPP and taxol MTs are well done. Credit is also due for the explicit paragraph saying they cannot rule out an alternative α sign for all protofilament numbers, which would invalidate their specific proposal.\n\nThe soft spots, in proportion. First, the experimental confirmation of handedness tunability depends on assigning α<0 and low κ̃ to taxol MTs; neither is measured. α from protofilament number is plausible, but the literature they cite includes the competing kinesin-sidestepping argument, and the κ̃ estimate for typical MTs (250) is right at the model's threshold (200). The paper concedes the alternative scenario and says taxol could simply be reclassified as high-κ̃. That is not a confirmation; it is a consistent story. Second, the anti-correlation between nematic order and rotation rate—their main evidence for collision-driven rotation—is confounded because the density also changed. Suggestive, but not clean. Third, their model in the high-κ̃ regime has no single-filament rotation, while their experiments show a small but significant CCW bias in isolated filaments; they argue it is too slow to matter, which is fair, but it is a real mismatch. Minor: Ref [29] is a lithium battery paper, clearly a citation error, and no code or data are shipped.\n\nThe simulation result alone is a real contribution and deserves careful refereeing. The experimental section should be revised to present the taxol result as a consistency check, or to actually measure the protofilament distribution and stiffness.\n\nI would take this to reading group, I would cite the simulation result, and I would send it to review.","headline":"Solid sufficiency result for chiral active forces in simulations; experimental confirmation of stiffness-tunable handedness rests on an admitted double-sign-flip that is not independently anchored.","tokens_in":26803,"tokens_out":5181,"would_cite":true,"duration_ms":56384,"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":"Chirality in the active forces motors exert on microtubules — not filament curvature — drives the coherent collective rotation seen in gliding assays, and filament stiffness sets its speed and handedness.","keywords":["microtubule gliding assay","chiral active forces","time cholesteric","active nematic order","filament bending rigidity","Brownian dynamics simulation","kinesin motor chirality","handedness propagation"],"falsifier":"A decisive check is a gliding assay made only of 13-protofilament microtubules, which have no protofilament skew and therefore no chiral force angle from that source: if the aligned state still rotates coherently, chiral active forces of the modeled kind are not the operative mechanism — the authors note this experiment is impractical with stabilized filaments. A second check is a continuous stiffness scan on a single microtubule type, for instance by titrating a crowding agent or crosslinker: the director's rotation should slow, stop, and reverse as the stiffness threshold near the model's $\\tilde{\\kappa}\\approx 200$ is crossed, and a handedness that never reverses with stiffness would falsify the tuning claim.","tokens_in":2347,"feed_emoji":"🔄","tokens_out":3592,"duration_ms":227618,"temperature":0.7,"pith_summary":"The paper proposes that the coherent rotation of the filament crowd in microtubule-kinesin gliding assays — a 'time cholesteric' state — comes from chirality in the active forces motors exert, not from curvature or shape chirality of the filaments. The authors model this as self-propulsion directed at a small fixed skew angle to the local filament tangent, and their Brownian dynamics simulations show that this single chiral ingredient is enough to produce rotating active nematic order (a long-range-aligned state whose average filament orientation steadily rotates), as long as colliding filaments can cross each other. The result is a tunable rotor: above a bending-rigidity threshold the director rotates with the same handedness as the molecular chirality, below it the rotation reverses, and at the threshold it stops. Experiments on glass and on lipid bilayers with two microtubule types (GMPCPP- and taxol-stabilized) support the collision-driven picture and a two-sign-flip account of why both rotate counterclockwise on lipids. If the mechanism is right, it supplies a minimal route for molecular handedness to reach the material scale and identifies filament stiffness as a control knob for chiral active matter.","feed_headline":"Motor chirality, not shape, rotates gliding microtubule crowds","feed_subtitle":"Filament collisions carry molecular handedness into the collective rotation, and stiffness can flip the spin direction.","key_machinery":"The carrying object is a two-dimensional Brownian dynamics model of self-propelled semiflexible filaments: each filament is a chain of 30 beads connected by Hookean springs, with a harmonic bending rigidity $\\tilde{\\kappa}$ that sets its stiffness, and every bead is driven by an active force of fixed magnitude oriented at a small skew angle $\\alpha$ to the local tangent — the only chiral ingredient in the model. Filament–filament interactions are shifted Weeks–Chandler–Andersen repulsions that let colliding filaments interpenetrate at a modest energy cost, which prior work found necessary for long-range nematic order. The argument's second mechanism is collision-mediated rotation: two-filament scattering simulations show that the bisector of a colliding pair rotates with a handedness set by $\\tilde{\\kappa}$, because a filament hindered at its head tends to rotate opposite to the chiral bias while one hindered at its tail rotates with it; aligning collisions dominate for flexible filaments and crossovers for rigid ones. An iterative Markov-chain weighting of the pair-collision statistics predicts the bulk handedness reversal from two-filament data alone.","core_discovery":"On the paper's own terms, the central claim is that an off-tangent component of the motor force on each microtubule segment — a chiral active force at angle $\\alpha$ to the local tangent — is sufficient to produce a bulk rotating nematic state in a gliding assay, with no need for intrinsic filament curvature. The supporting discovery is that collisions, rather than single-filament turning, transmit the chirality: the curvature measured on isolated filaments would predict rotation rates orders of magnitude below those observed, and lowering the filament density weakens nematic order while speeding up the director's rotation, as expected if collisions drive it. In the bead-spring simulation, a positive $\\alpha$ gives counterclockwise director rotation at high bending rigidity ($\\tilde{\\kappa}=300$) but clockwise rotation for $\\tilde{\\kappa}\\lesssim 200$, with the rotation vanishing at the threshold; pair-collision statistics reproduce the reversal by weighing aligning events (favored by flexible filaments, rotating against the molecular chirality) against crossovers (favored by rigid filaments, rotating with it). The authors also predict that chiral activity makes nematic order the stable long-range-ordered state for flexible filaments — reversing the achiral trend — and that some parameters admit bistable nematic and polar states rotating with opposite handedness.","pith_inferences":["Editorial inference — the head- versus tail-pinning picture implies a clean single-filament test the authors do not run: a microtubule blocked at its leading end by an obstacle should rotate opposite to one blocked at its tail, checkable in dilute assays with engineered barriers.","Editorial inference — if motor-driven chiral forces do contribute to cortical microtubule reorganization in plant cells as the authors speculate, their stiffness-tuned handedness predicts that interventions changing lattice rigidity (crosslinkers, acetylation, severing) should alter the sense or rate of array rotation.","Editorial inference — because the handedness reversal is tied to flexibility rather than to the sign of the chiral bias, synthetic chiral active rods pushed at a skew angle should show the same switch, making stiffness a general design parameter for rotor direction in active materials.","Editorial inference — the model's claim that collective rotation stabilizes nematic order implies that, across experimental conditions, the most persistently ordered nematic states should be those with the fastest director rotation; this correlation is measurable from existing assay movies."],"forward_implications":["Coherent collective rotation can arise in gliding assays without shape chirality, so observing rotation does not by itself license claims about intrinsic filament curvature.","Filament bending rigidity is a single-knob control for macroscopic chirality: crossing the stiffness threshold near $\\tilde{\\kappa}\\approx 200$ reverses the rotation's handedness, and at the threshold the rotation stops.","Collision frequency is the transmission channel: lower nematic order means more collisions and faster director rotation, so rotation rate and order parameter should be anti-correlated across conditions.","Chiral activity stabilizes nematic order over polar order in flexible-filament regimes, and produces initial-condition-dependent bistability in which nematic and polar states rotate with opposite handedness.","The same stiffness-controlled reversal is predicted for any chiral active system of penetrable semiflexible filaments driven at a fixed skew angle, a class the authors suggest extends beyond microtubule-kinesin assays."],"supporting_citations":[{"why":"Supplies the achiral semiflexible-filament simulation, the interpenetrable collision rule yielding nematic order, and the stiffness estimate $\\tilde{\\kappa}\\approx 250$ that this work extends with a chiral angle.","marker":"[33]"},{"why":"Prior report of large-scale counterclockwise rotation and the protofilament-count excesses used to assign the sign of the chiral angle in the two-sign-flip argument.","marker":"[16]"},{"why":"Prior gliding-assay study with global orientational order, clusters, and counterclockwise director rotation that motivates the collision picture.","marker":"[17]"},{"why":"Provides the lipid-bilayer mobile-motor assay yielding the uniform active nematic states on which both microtubule types were compared.","marker":"[26]"},{"why":"Proposed collision-induced torque for chiral patterns; the qualitative rotation-from-collisions effect the simulations reproduce without chiral interaction rules.","marker":"[35]"},{"why":"Shows kinesin-1 walks along the protofilament axis, justifying the active force angled to the filament tangent.","marker":"[20]"},{"why":"Reviews helical motor motion and torque generation, grounding the sign convention for the skew angle from protofilament supertwist.","marker":"[24]"},{"why":"Supplies the repulsive Weeks-Chandler-Andersen pair potential, shifted here so colliding filaments cross over at modest energy cost.","marker":"[43]"}],"fun_headline_variants":["Stiffness flips rotation direction in chiral microtubule crowds","Collisions transmit motor chirality to rotating microtubule nematic","Chiral motor forces rotate microtubule crowds without curvature","Bending stiffness tunes handedness of gliding microtubule rotation","Microtubule group spin driven by chiral motor forces, not shape"],"cache_read_input_tokens":28800,"weakest_assumption_plain":"The experimental interpretation hangs on the assumption that the sign of the effective chiral motor-force angle is fixed by the handedness of the helical protofilament twist — positive for GMPCPP microtubules, which have an excess of 14 protofilaments, and negative for taxol-stabilized ones, which have an excess of 12 — together with the assumption that taxol-stabilized microtubules sit in the low-stiffness regime where collective rotation reverses; the authors concede that if leftward motor sidestepping makes the effective angle positive for all protofilament numbers, their specific account of the taxol data would be invalidated.","fun_headline_variants_meta":{"raw":{"variants":["Stiffness flips rotation direction in chiral microtubule crowds","Collisions transmit motor chirality to rotating microtubule nematic","Chiral motor forces rotate microtubule crowds without curvature","Bending stiffness tunes handedness of gliding microtubule rotation","Microtubule group spin driven by chiral motor forces, not shape"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000177,"raw_usage":{"total_tokens":1331,"prompt_tokens":1018,"completion_tokens":313,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":634,"completion_tokens_details":{"reasoning_tokens":228}},"tokens_in":634,"tokens_out":313,"duration_ms":3847,"temperature":1.0,"reasoning_tokens":228,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:21:44.473096+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check is a gliding assay made only of 13-protofilament microtubules, which have no protofilament skew and therefore no chiral force angle from that source: if the aligned state still rotates coherently, chiral active forces of the modeled kind are not the operative mechanism — the authors note this experiment is impractical with stabilized filaments. A second check is a continuous stiffness scan on a single microtubule type, for instance by titrating a crowding agent or crosslinker: the director's rotation should slow, stop, and reverse as the stiffness threshold near the model's $\\tilde{\\kappa}\\approx 200$ is crossed, and a handedness that never reverses with stiffness would falsify the tuning claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the achiral semiflexible-filament simulation, the interpenetrable collision rule yielding nematic order, and the stiffness estimate $\\tilde{\\kappa}\\approx 250$ that this work extends with a chiral angle."},{"cited_title":"Large-scale chirality in an active layer of microtubules and kinesin motor proteins","cited_arxiv_id":null,"evidence_quote":"Prior report of large-scale counterclockwise rotation and the protofilament-count excesses used to assign the sign of the chiral angle in the two-sign-flip argument."},{"cited_title":"Gliding filament system giving both global orientational order and clusters in collective motion.Physical Review E, 101(3), March 2020","cited_arxiv_id":null,"evidence_quote":"Prior gliding-assay study with global orientational order, clusters, and counterclockwise director rotation that motivates the collision picture."},{"cited_title":"Memarian, Joseph D","cited_arxiv_id":null,"evidence_quote":"Provides the lipid-bilayer mobile-motor assay yielding the uniform active nematic states on which both microtubule types were compared."},{"cited_title":"Rashedul Kabir, and Akira Kakugo","cited_arxiv_id":null,"evidence_quote":"Proposed collision-induced torque for chiral patterns; the qualitative rotation-from-collisions effect the simulations reproduce without chiral interaction rules."},{"cited_title":"Kinesin follows the microtubule’s protofilament axis.Journal of Cell Biology, 121(5):1083–1093, June 1993","cited_arxiv_id":null,"evidence_quote":"Shows kinesin-1 walks along the protofilament axis, justifying the active force angled to the filament tangent."},{"cited_title":"Helical motion and torque generation by microtubule motors.Current Opinion in Cell Biology, 88:102367, June 2024","cited_arxiv_id":null,"evidence_quote":"Reviews helical motor motion and torque generation, grounding the sign convention for the skew angle from protofilament supertwist."},{"cited_title":"Weeks, David Chandler, and Hans C","cited_arxiv_id":null,"evidence_quote":"Supplies the repulsive Weeks-Chandler-Andersen pair potential, shifted here so colliding filaments cross over at modest energy cost."}],"review_version":1}