{"id":"3b924072-5d2e-4ec8-9d56-b36e6d9df1f9","arxiv_id":"2504.20893","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Simulations of E. coli with flexible polymorphic flagella show that tumbling near a wall produces smaller tumble angles and a strong forward bias in in-plane reorientation.","lead":"A detailed simulation of E. coli tumbling shows that near a solid surface bacteria reorient far less than in open water, often keeping almost the same heading. This may explain why experiments report fewer tumble events near surfaces and why bacteria appear to run longer there.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Near-wall tumble statistics may be conditional on which flagellum reverses: the wall breaks the fourfold symmetry the model relies on in bulk, and Section 4.2 never states whether all four motors were sampled.","rationale":"I focused on the sampling of flagellum identity because the central claim is a marginal statement about E. coli tumbling, while the simulation design deliberately fixes the reversing flagellum for tractability (Section 2.2). In bulk this is harmless; near a wall it is not, because the wall makes the four attachment points inequivalent. The paper does not document the body-fixed flagellum used, the initial roll angle, or any averaging over the four motors, so the 25% and mean-41° numbers cannot currently be assessed as biological averages. This is a gap in the mapping from simulation ensemble to experimental ensemble, not a disagreement with consensus. The reader's Reynolds-number concern is legitimate and would also benefit from a convergence test, but it applies to the whole method and is partially mitigated by the bulk tumble-angle agreement with Berg and Brown; the flagellum-selection issue is specific to the new near-wall result and can be resolved with modest additional simulations. I keep the reader's CONDITIONAL verdict: the paper is plausible and internally consistent, but the near-wall quantitative claims need this condition stated and tested.","tokens_in":17885,"tokens_out":9286,"duration_ms":93733,"concrete_test":"Repeat the Section 4.2 single-tumble ensemble (τt = 30τb, same steady wall distance) four times, reversing a different one of the four flagella in each run, first with identical initial roll angle and then with randomized initial roll. Recompute ⟨γ⟩, P(φ), P(θ), and the fraction φ < 10° for each subset and for the equally weighted mixture. If the fraction φ < 10° varies across the four flagella by more than the binomial sampling error, the headline statistic is conditional on motor identity; report the marginal distribution. Also record the flagellum-wall separation at tumble initiation to verify whether the setup already biases which flagellum interacts with the wall.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative near-wall claim—mean tumble angle 41° and 25% of tumbles with in-plane angle below 10° (Section 4.2, Fig. 8)—is presented as a property of E. coli tumbling near a surface. The model has four flagella symmetrically attached to the rear (Section 2.2), so in bulk the choice of reversing flagellum is irrelevant. Near a wall the symmetry is broken: one flagellum is wall-proximal, one distal, two lateral. Section 4.2 states only that 'tumbling is initiated by reversing the rotation of one flagellum' and does not report which flagellum, whether the initial roll angle around the body axis was randomized, or whether all four flagella were simulated. If the same body-fixed flagellum was reversed in every single-tumble run, the reported P(γ), P(φ), and the 25% forward-bias fraction are conditional distributions for one motor position. Experiments observe stochastic reversals across flagella, so the marginal distribution over the four motors is the quantity needed to support the proposed explanation of suppressed tumbling and prolonged runs. The long trajectory (Fig. 7) contains only eight near-wall tumble events and cannot by itself average over motor identity.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a mesoscale MPCD model of E. coli with four flexible flagella modeled by an extended Kirchhoff-rod theory that includes polymorphic transformations, and uses it to study single tumble events in bulk and near a flat no-slip wall. In bulk, for a fixed tumble time of 30 bundle periods the simulated tumble-angle distribution (mean 61°, standard deviation 34°) is compared with Berg and Brown's experimental values (mean 62°, standard deviation 26°); increasing the hook bending rigidity narrows the distribution and reduces flagellar dispersion. Near a wall, independent tumble simulations yield a mean tumble angle of 41° (versus 61° in bulk), an in-plane reorientation distribution in which 25% of events have an in-plane angle below 10°, and a polar-angle distribution peaking near 90°. The authors propose that these small in-plane reorientations explain the experimentally reported suppression of tumbling and prolonged run times near surfaces.","tokens_in":18141,"tokens_out":6610,"duration_ms":69395,"significance":"If the near-wall statistics are robust, this is a useful mechanistic hypothesis for why tumbling near surfaces appears suppressed: tumbles still occur but reorient the cell mostly within the surface plane and often with very small heading change. The model's strengths are its explicit treatment of flagellar polymorphism, flexible hooks, and hydrodynamic coupling, and the bulk tumble-angle statistics are based on 932–1032 independent events and agree well with a classic experiment. The near-wall predictions, however, rest on two load-bearing assumptions that are not yet verified: the fluid dynamics is computed at Reynolds number about 0.4 rather than the Stokes regime of real E. coli, and the near-wall tumble statistics may be conditional on a single flagellar motor position because of the assumed fourfold axisymmetry. These issues, rather than the bulk phenomenology, determine whether the quantitative claims (41° mean angle, 25% forward bias) can be transferred to real bacteria.","major_comments":[{"comment":"Section 2.4 reports Re≈0.4 for the MPCD fluid, roughly four orders of magnitude above the value for E. coli. The near-surface effects in Fig. 8 — reduced tumble angle, forward-biased in-plane reorientation, and reduced flagellar dispersion — are attributed to wall-induced hydrodynamic interactions and enhanced drag. At Re=0.4 inertial contributions are not negligible, so the quantitative distributions may not represent the Stokes regime that governs real E. coli. Please either perform a convergence study at lower Re (for example by increasing the viscosity or reducing the lattice constant) and show that P(γ), P(φ), and ⟨I1⟩/I_bundle^1 are unchanged, or provide a quantitative argument that the wall-induced flows relevant to tumbling are Stokes-like despite the nominal Reynolds number.","section":"Section 2.4 and Section 4.2"},{"comment":"Section 2.2 states that the bacterium is 'fully axisymmetric' so that the authors can avoid 'randomly choosing the specific position of the reversely rotating flagellum.' This is a deliberate modeling choice, but it makes the bulk and near-wall cases different: in bulk the four flagella are equivalent, while near a wall one flagellum is closest to the surface and the others are not. Section 4.2 says only that 'tumbling is initiated by reversing the rotation of one flagellum' and does not report whether all four motor positions were sampled, whether the initial roll angle about the body axis was randomized, or whether the same flagellum was used throughout. If a single body-fixed flagellum was used, the 41° mean and the 25% below-10° fraction in Fig. 8 are conditional distributions, not the marginal distributions that experiments observe where motor reversals are stochastic across flagella. The eight near-wall events in Fig. 7 are too few to average over motor identity. Please report the sampling protocol and, if only one motor was used, provide the marginal P(γ), P(φ), and P(θ) over the four motors or justify that the conditional statistics are representative.","section":"Section 2.2 and Section 4.2"},{"comment":"Unlike the bulk case, where the text reports N=932 and N=1032 for the two tumble-time protocols, Section 4.2 does not state the number of independent near-wall tumble events used for Fig. 8, and no error bars are shown. The headline numbers — mean γ=41°, 25% of events with φ<10°, and 53% escape-oriented events — need statistical support. Please report N and uncertainties (for example, bootstrap confidence intervals) for all three distributions.","section":"Section 4.2"}],"minor_comments":[{"comment":"Equation (3) includes a term A/2 ξ²(∂sΩ)² that is said to allow smooth transitions between polymorphic regions of size ξ, but the numerical value of ξ is never given; please state it in Section 2.4.","section":"Section 2.1"},{"comment":"The text says that for hook rigidities of 100Ah and 400Ah the tumble-angle distributions 'become more narrow,' but the reported mean angles of 87° and 69° are above the reference mean of 61°; please clarify whether 'narrow' refers to the variance and report the standard deviations for these distributions.","section":"Section 3.3, Fig. 5"},{"comment":"The model relies on several phenomenological parameters (ground-state energies δ(n), hook rigidity Ah, motor reversal protocol) and the paper does not assess sensitivity of the near-wall results to these choices; a brief sensitivity discussion would strengthen the conclusions.","section":"Section 5 and Data availability"},{"comment":"The data availability statement says data are available upon request from one of the authors; for reproducibility, please consider depositing the simulation code and parameter files in a public repository.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague —\n\nThe paper deserves a serious look. Its real new result is the near-wall tumble statistics: compared to bulk (mean 61°), tumbling near a wall shifts the tumble-angle distribution down (mean 41°) and produces a strong in-plane forward bias, with 25% of events below 10°. That is a concrete, mechanistic way to reconcile ref. 39's report that tumbling is suppressed at surfaces with ref. 41's claim that it drives escape. The result is an emergent output of the simulation, not a fit, so it has predictive weight.\n\nWhat the paper does well: the model is the group's established Kirchhoff-rod/MPCD framework, extended to tumbling near a surface; they validate against Berg-Brown in bulk and get a decent match for a fixed tumble time. The hook-stiffness sweep and the dispersion measure (smallest eigenvalue of the bundle's moment of inertia) are thoughtful and clearly explained. Also to their credit, they explicitly flag the four-flagella symmetry as a simplification and say the influence on tumbling needs further work—that is honest.\n\nSoft spots, in rough order of importance. First, the Reynolds number: Re≈0.4 is four orders of magnitude above real E. coli, and the near-wall results depend on wall-induced hydrodynamics. They report that the swimming speed and bundle rotation are off from experiments (though the ratio v/ωr is right), and they don't test convergence toward Stokes flow. That is a real gap, though arguably a known MPCD limitation. Second, the paper is silent on which flagellum reverses in the near-wall runs. Section 2.2 says the symmetric design lets them avoid randomizing the reversing flagellum; that is valid in bulk, but the wall breaks the symmetry. If all near-wall tumble events used the same body-fixed flagellum, the P(γ), P(φ), and the 25% figure are conditional on motor position, not the marginal distribution an experiment would sample. The long trajectory has only eight near-wall events and can't average over that either. This is not fatal—the effect might survive averaging—but the authors need to state the protocol and ideally run all four motors. Third, the phenomenological ground-state energies are hand-tuned to produce curly-I; the paper leans on earlier work for that, which is fine but leaves the model sensitive to parameters that aren't independently fixed. Fourth, no error bars on the quoted fractions, and data are only 'available upon request.'\n\nWho benefits: anyone working on bacterial motility, surface residence, or flagellar polymorphism; the near-wall reorientation result is worth discussing even with the caveats. I'd accept it for peer review and ask the authors to clarify the motor-identity protocol, add a low-Re check if feasible, and provide error bars or raw data.","headline":"Near-wall tumbling simulation with a genuinely new in-plane forward-bias result; solid bulk validation, but authors must clarify which flagellum reverses and address the Re=0.4 gap.","tokens_in":18696,"tokens_out":2469,"would_cite":true,"duration_ms":24743,"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":"Near a bounding surface, E. coli tumble angles shrink and turn forward-biased, explaining reported suppression of tumbling.","keywords":["E. coli","tumbling","flagellar polymorphism","Kirchhoff rod","multiparticle collision dynamics","surface hydrodynamics","run-and-tumble","tumble angle distribution"],"falsifier":"Repeat the near-wall tumble simulations with the same bacterium at Reynolds number 0.4 and at about $10^{-5}$ (for example, by raising viscosity or lowering motor torque). If the mean tumble angle near the wall does not stay well below the bulk value of 61 degrees, or if the fraction of in-plane angles below 10 degrees drops from 25%, the central claim fails. On the experimental side, a high-frame-rate (at least 200 Hz) tracking study of individual E. coli near a glass surface that counts every velocity drop would settle whether such small in-plane tumbles actually occur.","tokens_in":17683,"feed_emoji":"🦠","tokens_out":4896,"duration_ms":47193,"temperature":0.7,"pith_summary":"The paper argues that the surface itself reshapes how E. coli tumbles: close to a flat wall, the tumble angle distribution shifts to smaller angles, flagellar dispersion drops, and in-plane reorientation becomes strongly forward-biased, with 25% of near-wall tumbles turning less than 10 degrees. This provides a mechanistic explanation for why earlier experiments reported suppressed tumbling near surfaces: many near-wall tumbles are simply too small to notice. The same model first reproduces the classic bulk tumble-angle distribution, giving confidence that the near-surface changes are caused by the wall rather than by model artifacts. A sympathetic reader would care because it connects single-flagellum mechanics to surface exploration, biofilm formation, and bacterial escape.","feed_headline":"Near a wall, E. coli tumbles favor small forward turns","feed_subtitle":"Simulation finds 25 percent of near-wall tumbles reorient less than 10 degrees in-plane, explaining reports of suppressed tumbling.","key_machinery":"The central object is a discrete flagellum built from an extended Kirchhoff-rod elastic energy, in which each segment's rotational strain vector chooses among four polymorphic ground states (normal, coiled, semi-coiled, curly-I) with a smooth transition cost. The flagella are hydrodynamically coupled to the fluid through multiparticle collision dynamics, which supplies the wall-induced flows and drag. The hook is represented by reduced bending rigidity at the motor attachment. This machinery is what lets a single reversal of one motor produce realistic polymorphic transformation, bundle unbundling, and reorientation statistics, both in bulk and at the wall.","core_discovery":"In a simulation of E. coli with four flexible, polymorphic flagella in an MPCD fluid, a tumble event is triggered by reversing one motor. In bulk, the tumble-angle distribution has mean 61 degrees and matches the classic experiment (mean 62 degrees). Near a no-slip surface, the mean drops to 41 degrees, the polar angle peaks at 90 degrees (parallel to the wall), and the in-plane angle distribution is forward-biased: 25% of events have in-plane reorientation below 10 degrees. The authors claim these small forward-biased events are the origin of the experimentally reported 'suppressed' tumbling near surfaces, and that they help bacteria stay at the surface while still exploring it.","pith_inferences":["If the forward-bias is real, then the standard run-and-tumble picture near surfaces should be revised: runs may not be longer, but each tumble changes direction less, so effective exploration is dominated by many small turns.","A testable extension: track near-wall tumbles with high-speed microscopy and count all velocity drops; the model predicts a population of tumbles with in-plane angle below 10 degrees that previous assays may have discarded.","Because the simulation runs at Reynolds number about 0.4, roughly 10^4 times the real value, the near-wall forward bias should be re-checked at lower Reynolds numbers before it is used to interpret experiments.","The same polymorphic flagellum model could be applied to other peritrichous bacteria to see whether surface-biased tumbling is generic or specific to E. coli."],"forward_implications":["If the wall shifts tumbling to small forward angles, near-wall E. coli effectively reorient in place, which should increase their residence time at surfaces without requiring fewer tumble events.","Stiffer hooks narrow the tumble-angle distribution and reduce flagellar dispersion; in the model, a 100-fold stiffer hook raises the mean tumble angle and eliminates very large angles.","Because 53% of near-wall tumbles give a polar angle below 90 degrees, tumbling does provide escape attempts, but escape is not simply correlated with flagellar dispersion.","The measured bulk agreement pins the model's tumble mechanics to real E. coli, so the surface comparison is the meaningful new quantity."],"supporting_citations":[{"why":"Supplies the classic experimental tumble-angle distribution (mean 62 degrees) that the bulk simulation matches.","marker":"[14]"},{"why":"The experimental report of reduced tumble frequency and small reorientation near surfaces that the paper's forward-bias result explains.","marker":"[39]"},{"why":"Provides the Gamma-distribution tumble times, behavioral-variability discussion, and the claim that tumbling efficiently escapes surfaces.","marker":"[41]"},{"why":"Establishes the hydrodynamic pusher attraction that traps E. coli near surfaces, the context for near-wall tumbling.","marker":"[17]"},{"why":"Source for motor torque, bundle rotation frequency, and polymorphic transformation parameters used in the model.","marker":"[13]"},{"why":"Earlier polymorphic flagellum model with Kirchhoff-rod polymorphism that this work extends.","marker":"[21]"},{"why":"Supplies the ground-state rotational strain vectors for the four polymorphic states.","marker":"[55]"},{"why":"Prior simulation of E. coli near surfaces that the long-trajectory collision behavior is compared against.","marker":"[35]"}],"fun_headline_variants":["Wall proximity shrinks E. coli tumble angles in simulations","Simulation: near-wall E. coli tumbles bias to small forward turns","E. coli near surfaces: tumbling prefers small forward reorientations","Surface-adjacent E. coli tumble with narrower, forward-biased angles","In MPCD simulation, wall biases E. coli tumble to small turns"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The near-wall results assume that the simulated fluid at Reynolds number about 0.4 behaves like real E. coli hydrodynamics at Reynolds number about $10^{-5}$; if inertial effects at 0.4 change the wall-induced drag on flagella, the reduced tumble angles and forward bias could be simulation artifacts.","fun_headline_variants_meta":{"raw":{"variants":["Wall proximity shrinks E. coli tumble angles in simulations","Simulation: near-wall E. coli tumbles bias to small forward turns","E. coli near surfaces: tumbling prefers small forward reorientations","Surface-adjacent E. coli tumble with narrower, forward-biased angles","In MPCD simulation, wall biases E. coli tumble to small turns"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000236,"raw_usage":{"total_tokens":1481,"prompt_tokens":900,"completion_tokens":581,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":516,"completion_tokens_details":{"reasoning_tokens":483}},"tokens_in":516,"tokens_out":581,"duration_ms":5823,"temperature":1.0,"reasoning_tokens":483,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:17:57.693427+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the near-wall tumble simulations with the same bacterium at Reynolds number 0.4 and at about $10^{-5}$ (for example, by raising viscosity or lowering motor torque). If the mean tumble angle near the wall does not stay well below the bulk value of 61 degrees, or if the fraction of in-plane angles below 10 degrees drops from 25%, the central claim fails. On the experimental side, a high-frame-rate (at least 200 Hz) tracking study of individual E. coli near a glass surface that counts every velocity drop would settle whether such small in-plane tumbles actually occur.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the classic experimental tumble-angle distribution (mean 62 degrees) that the bulk simulation matches."},{"cited_title":"Molaei, M","cited_arxiv_id":null,"evidence_quote":"The experimental report of reduced tumble frequency and small reorientation near surfaces that the paper's forward-bias result explains."},{"cited_title":"Junot, T","cited_arxiv_id":null,"evidence_quote":"Provides the Gamma-distribution tumble times, behavioral-variability discussion, and the claim that tumbling efficiently escapes surfaces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the hydrodynamic pusher attraction that traps E. coli near surfaces, the context for near-wall tumbling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source for motor torque, bundle rotation frequency, and polymorphic transformation parameters used in the model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier polymorphic flagellum model with Kirchhoff-rod polymorphism that this work extends."},{"cited_title":"Vogel and H","cited_arxiv_id":null,"evidence_quote":"Supplies the ground-state rotational strain vectors for the four polymorphic states."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior simulation of E. coli near surfaces that the long-trajectory collision behavior is compared against."}],"review_version":1}