{"id":"36dd001e-834b-421c-8d02-81a2f4d72633","arxiv_id":"1908.03597","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"For two interacting molecules searching a 1D lattice, the mean search time is non-monotonic in the dimer dissociation rate, with the fastest search at an intermediate interaction strength.","lead":"A model of two proteins searching DNA found that when they can form a temporary pair, the search is fastest at an intermediate pairing strength, not at the strongest or weakest. The result suggests cells might tune protein-protein interactions to speed up gene activation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Simulation protocol for the headline non-monotonic result never specifies the association rate ka or any rule for dimer formation, leaving Fig. 3 unreproducible and the central claim conditional on an unstated implementation.","rationale":"The paper's key contribution is the prediction, based on Monte Carlo simulations, that tuning the dimer dissociation rate kd can produce a non-monotonic search time with a minimum at intermediate interaction strengths. This is 'one of our main results' and is showcased in Fig. 3. The analytical sections cover only the two limits (independent monomers and stable dimer); the intermediate regime—where the minimum appears—has no analytic treatment and is supported solely by the simulations. The simulation method paragraph lists five parameters but then describes only monomer hopping and dimer dissociation/movement. Association is never mentioned, and ka is never assigned a value. This is not a stylistic imperfection: the relative magnitude of ka controls whether a complex forms at all when two monomers meet, and the model's own limits are phrased in terms of ka versus µ1. If the implemented code used ka=∞, the result is a two-parameter model; if a finite ka was used, the curves in Fig. 3 could change. Either way, the published description does not permit reproduction, and the non-monotonic claim cannot be independently checked. The reader's weakest_assumption identified the same ka gap together with the µ2<µ1 ordering; I focus on the ka gap because it is closer to internal reproducibility than to a physical assumption. A deeper worry—that the minimum could be a finite-time-step or finite-sampling artifact—is secondary but reinforces the need for the authors to document and, ideally, provide error bars. I therefore keep the reader's CONDITIONAL verdict unchanged.","tokens_in":8923,"tokens_out":17964,"duration_ms":168558,"concrete_test":"Contact the authors for the exact value of ka and the association algorithm used in the Fig. 3 simulations, then independently rerun the same protocol for ka/µ1 = 0.1, 1, 10, and ∞ (instantaneous association). If the non-monotonic minimum persists for all ka ≥ µ1, the result is robust; if the minimum disappears or moves for finite ka, the headline claim depends on an unstated parameter. A separate analytical check would be to derive a two-state (monomer/dimer) approximation for the MFPT in the 1D+2D regime and verify that it predicts a minimum in kd for µ2 < µ1; if no such minimum exists for any finite ka, the simulation result is suspect.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that the mean first-passage time is non-monotonic in the dissociation rate kd, with a minimum at intermediate intermolecular interactions—rests entirely on the general-regime Monte Carlo simulations in Fig. 3. Section II ('Simulation method') states that the model has five kinetic parameters (µ1, µ2, ka, kd, koff), sets µ1 = 0.25, and then lists the moves: a monomer remains with probability 1/2 or hops left/right with probability 1/4; a complex dissociates with probability kd∆t, moves with probability µ2∆t, or remains with probability 1−kd∆t−µ2∆t. No attempt is made to associate two monomers found on the same site, and no value for ka is ever given. Since the model distinguishes the limits ka≪µ1 (no complex) from ka≫µ1 (instantaneous complex formation), the simulation must include some association event, but the description omits it entirely. If the implementation used instantaneous association upon contact (ka=∞), the observed non-monotonicity might still hold for the stated µ2<µ1; but if a finite ka was used, the minimum in Fig. 3 could shift or vanish depending on the chosen value. The paper's own detailed-balance relation ka/kd = exp(−E/kBT) ties ka to the interaction energy E, so leaving ka unspecified is not a minor omission: the 'intermediate interaction' regime is precisely the range where both monomers and dimers coexist, and that coexistence is governed by the ratio ka/kd. Without the actual ka, the reader cannot reproduce Fig. 3, nor determine whether the claimed optimal interaction strength has any physical meaning.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a minimal one-dimensional lattice model in which two interacting molecules search for a target site on DNA. The molecules can reversibly associate into a dimer that also diffuses along the lattice, and the search is completed when both molecules are simultaneously present at the target. The authors analyze strong-repulsion and strong-attraction limits with first-passage formulas and perform kinetic Monte Carlo simulations for the general regime. Their central claim is that the mean first-passage time is non-monotonic in the dimer dissociation rate kd, so that an intermediate interaction strength gives the fastest target search. They also study the effects of target residence time and target position, and they interpret the intermediate regime as a beneficial combination of effectively one-dimensional and two-dimensional search mechanisms.","tokens_in":9312,"tokens_out":8120,"duration_ms":88579,"significance":"If the central claim is correct, the paper addresses a real gap in the protein target-search literature, where intermolecular interactions are usually neglected and a non-monotonic optimum in interaction strength would be a nontrivial, potentially testable prediction. The limiting-case analytical expressions, such as the L^2 scaling for two independent particles and the approximate complex-search formula, are transparent and are shown to agree with simulations. The physical interpretation via a 1D+2D trade-off is plausible and connects to facilitated-diffusion ideas. However, the headline non-monotonic result currently rests on simulations whose association protocol is not specified and that are shown without statistical error bars, so the significance is conditional on those simulations being reproducible and statistically robust.","major_comments":[{"comment":"The simulation protocol is incomplete because no association event is specified. The model definition includes the association rate ka and the detailed-balance relation ka/kd = exp(-E/kBT), but the described Monte Carlo algorithm lists only monomer moves, dimer dissociation/moves, monomer unbinding, and the dissociation event; nothing states what happens when two monomers occupy the same site, with what probability per time step, or what value of ka was used for the reported curves. Since the general-regime data in Fig. 3 depend on the coexistence of monomers and dimers, which is governed by ka/kd, the central non-monotonic claim cannot currently be reproduced or independently assessed. The authors should add the association update rule (including how simultaneous arrivals are treated), the ka values used for each figure, and a check that the limits ka >> mu1 and ka << mu1 are realized as claimed.","section":"Section II, Simulation method"},{"comment":"The headline non-monotonic dependence of the mean first-passage time on the dissociation rate kd is presented as smooth curves without error bars, numbers of Monte Carlo runs, or raw data points. The dip at intermediate kd, especially for mu2/mu1 = 0.2, may be modest compared with stochastic fluctuations, and the reader cannot determine whether the reported minimum is statistically significant. Please add standard errors or other uncertainty measures, report the number of independent runs, and preferably show individual simulation points on the plot.","section":"Fig. 3 and Section III.C"},{"comment":"The additional general-regime results in Figs. 4 and 5 are generated with the same underspecified association protocol. Once the association rule and ka values are supplied, these figures should be regenerated or at least explicitly linked to the specified parameter set. Without this, the heat map in Fig. 5 and the trajectory interpretation in Fig. 4 inherit the reproducibility problem of the central simulation claim.","section":"Section III.C and Figs. 4-5"}],"minor_comments":[{"comment":"The text 'two articles move independently' appears to be a typo for 'two particles move independently'.","section":"Section III.C"},{"comment":"The axes labeled 'Log[kd]' and 'Log[koff]' do not specify the logarithm base, and the color-bar values have no units; using for example 'log10(kd/mu1)' would improve clarity.","section":"Fig. 5 caption and axes"},{"comment":"The assumption mu2 < mu1 is stated as a general expectation, and the non-monotonic optimum is demonstrated only in this regime. It would be useful to state explicitly that the predicted intermediate optimum is conditioned on mu2 < mu1 and to add a sentence or a simulation showing what happens when mu2 >= mu1.","section":"Section II"},{"comment":"The exponential approximation for the single-particle survival probability is an assumption; Fig. 2 shows good agreement for the integrated quantities used, but a brief comment on the range of lattice sizes and time scales for which this approximation is reliable would be helpful.","section":"Section III.A, Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is the missing association step in the simulation method, which undermines reproducibility of the central non-monotonic result. This is fixable with a clear algorithmic description and explicit ka values, so I do not recommend rejection. I also suggest requesting error bars in Fig. 3, as the magnitude of the dip relative to statistical noise is central to the paper's message."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the idea is genuinely nice and the two analytic limits check out, but the paper's headline result—the non-monotonic search time in the dimer dissociation rate—rests entirely on Monte Carlo simulations whose association step is never described. That needs fixing before I'd trust the claim.\n\nWhat's new: prior multi-particle first-passage papers (Grebenkov; Lawley and Madrid) treat non-interacting particles. Adding reversible dimer formation and showing a 1D+2D mixed regime with an optimal interaction strength is a real extension. The strong-repulsion and strong-attraction limits are clean, and the match with simulations there is good. The physical explanation (slow dimer vs. two fast monomers, with an optimum balance) is plausible and presented honestly.\n\nSoft spots: the simulation method in Section II lists five parameters but never specifies ka or any rule for two monomers forming a complex. The monomer update has three outcomes (stay, hop left, hop right); the complex update has dissociation and hop. There is no 'if two monomers occupy the same site, associate with probability X' anywhere. Since the intermediate regime is exactly where monomers and dimers coexist, the ka/kd ratio controls that coexistence, so leaving ka out is not a cosmetic omission. Fig. 3 plots MFPT vs kd at fixed (unspecified) ka; without the actual value and update rule, the non-monotonic minimum cannot be reproduced, and the claimed 'optimal interaction strength' has no quantitative meaning. The stress-test note is on target. Also, no error bars on the non-monotonic points, but that's secondary.\n\nThe analytic sections are fine and cite standard formulas correctly. The model's own limitations paragraph is honest about biological oversimplifications but doesn't mention the missing simulation detail—unfortunate.\n\nWho's this for: people working on facilitated diffusion and multi-particle first-passage problems in biophysics. The idea is worth pursuing; the execution is incomplete as submitted. I'd send it to referees, but instruct them to require a complete simulation protocol and a sensitivity check on ka before publication.","headline":"Nice idea and clean analytic limits, but the headline non-monotonic result is built on simulations whose association step is never specified; needs revision before I'd trust it.","tokens_in":9792,"tokens_out":2069,"would_cite":false,"duration_ms":22581,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Two interacting molecules can find a DNA target fastest at an intermediate dimer-dissociation rate, rather than as fully bound or fully independent particles.","keywords":["protein-DNA target search","first-passage time","dimer formation","protein-protein interaction","facilitated diffusion","kinetic Monte Carlo","gene regulation","random walk"],"falsifier":"Repeat the simulations behind Fig. 3 with the dimer hopping rate $\\mu_2$ larger than the monomer rate $\\mu_1$ while keeping all other rates fixed: if the mean first-passage time versus $k_d$ still shows a minimum, the proposed slow-dimer mechanism cannot be the cause. A cleaner test is a single scan with $\\mu_2/\\mu_1 > 1$; the paper’s interpretation predicts the mixed-regime minimum should disappear.","tokens_in":8723,"feed_emoji":"🧬","tokens_out":8727,"duration_ms":87080,"temperature":0.7,"pith_summary":"The paper asks how quickly two interacting protein molecules can find the same site on DNA, the step that precedes gene activation or repression. It models the pair as monomers hopping on a one-dimensional lattice who can meet and reversibly form a dimer complex, and the search ends only when both molecules occupy the target at the same time. The central claim is that the mean search time is not monotonic in the dimer dissociation rate: for intermediate interaction strengths, the search can be faster than either the fully independent or fully complexed extremes, because the system mixes a one-dimensional dimer search with a two-dimensional independent search. Tuning protein-protein interaction strength is therefore a plausible biological control knob for the speed of gene regulation.","feed_headline":"Dimer that sometimes splits finds DNA target fastest","feed_subtitle":"A two-protein model shows a mid-strength bond can beat both strong and weak sticking.","key_machinery":"The carrying object is the reversible dimer complex: two monomers hop with rate $\\mu_1$, associate at rate $k_a$ when co-located, dissociate at rate $k_d$, and the complex itself hops with rate $\\mu_2 < \\mu_1$. Detailed balance ties $k_a/k_d$ to the interaction energy $E$, so tuning $k_d$ is physically tuning the strength of protein-protein attraction. The target site has an unbinding rate $k_{\\rm off}$, and the search completes only when both molecules are on the target simultaneously. The argument combines first-passage survival-probability calculations in the non-interacting and strong-attraction limits with kinetic Monte Carlo simulations for the general regime.","core_discovery":"For two molecules that reversibly form a complex on a one-dimensional DNA segment, the mean first-passage time to simultaneous target occupancy is non-monotonic as a function of the dimer dissociation rate $k_d$, provided the dimer moves more slowly than the monomers. At large $k_d$ the molecules act independently and the search is effectively two-dimensional; at small $k_d$ the search is carried by a long-lived dimer and is effectively one-dimensional. In between, a mixed “1D+2D” regime can be fastest, and the paper explains this by an optimal balance between the dimer’s slow but directed progress and the monomers’ faster but wider exploration. The paper also finds that longer target residence times (smaller $k_{\\rm off}$) accelerate the two-molecule search and that targets away from lattice boundaries are reached sooner, with the non-monotonic dependence most visible for short residence times.","pith_inferences":["Because $k_a/k_d = \\exp(-E/k_B T)$, the optimal $k_d$ corresponds to a finite window of binding free energy; mutations that shift $E$ by a few $k_B T$ could move a pair between fast and slow regimes.","With more than two molecules, the same machinery would add another encounter bottleneck, so the optimal dissociation rate may shift or the non-monotonic minimum may sharpen; the paper’s construction can be extended to $N>2$ by simulation.","Real searches include 3D excursions; replacing the one-dimensional lattice by an intermittent 3D+1D walk should preserve the qualitative minimum but change its location, which could be tested in stretched-DNA or single-molecule experiments with two labelled proteins."],"forward_implications":["If the mixed regime is fastest, the cell can accelerate gene activation by tuning transcription-factor dimerization affinity to an intermediate value rather than maximizing or minimizing protein-protein binding.","Longer target residence times for a single molecule speed up the two-molecule completion, so proteins that bind their target strongly once they arrive reduce the time the other molecule needs to find it.","Targets located away from the segment boundaries are found faster, and the effect persists across interaction strengths.","The three search regimes imply distinct trajectory signatures: diagonal-bound motion for dimers, broad independent exploration for repulsion, and intermittent diagonal excursions for the mixed regime.","The non-monotonic optimum appears at short target residence times, so experiments with fast-dissociating transcription factors are the likeliest place to see it."],"supporting_citations":[{"why":"Provides the analytical treatment of two independent particles with reversible target-binding kinetics that the non-interacting limit builds on.","marker":"[11]"},{"why":"Supplies the kinetic Monte Carlo simulation method used for the general-regime results, including the non-monotonic curve.","marker":"[16]"},{"why":"Gives the single-particle mean first-passage time formulas for a 1D random walk on an interval used in the non-interacting and strong-attraction limits.","marker":"[31]"},{"why":"Supplies the intermittent-search concept used to explain why mixing 1D and 2D search modes can be faster than either alone.","marker":"[5]"},{"why":"Provides the first-passage results for random walks in bounded domains used for the $L^2 \\ln L$ scaling when the target acts as a reflecting boundary.","marker":"[7]"}],"fun_headline_variants":["Moderate dimer bond beats strong or weak for DNA search","Search time dips at optimal dimer splitting rate","Mixed 1D+2D search can be fastest for two proteins","Two-protein target search non-monotonic in dimer dissociation","Right balance of dimer splitting speeds DNA target search"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The non-monotonic optimum assumes that a two-molecule complex moves more slowly along DNA than a single molecule does; if that ordering is reversed, the mixed-search speed-up predicted here would not hold.","fun_headline_variants_meta":{"raw":{"variants":["Moderate dimer bond beats strong or weak for DNA search","Search time dips at optimal dimer splitting rate","Mixed 1D+2D search can be fastest for two proteins","Two-protein target search non-monotonic in dimer dissociation","Right balance of dimer splitting speeds DNA target search"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000287,"raw_usage":{"total_tokens":1678,"prompt_tokens":927,"completion_tokens":751,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":672}},"tokens_in":543,"tokens_out":751,"duration_ms":8229,"temperature":1.0,"reasoning_tokens":672,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:08:55.306377+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the simulations behind Fig. 3 with the dimer hopping rate $\\mu_2$ larger than the monomer rate $\\mu_1$ while keeping all other rates fixed: if the mean first-passage time versus $k_d$ still shows a minimum, the proposed slow-dimer mechanism cannot be the cause. A cleaner test is a single scan with $\\mu_2/\\mu_1 > 1$; the paper’s interpretation predicts the mixed-regime minimum should disappear.","supporting_citations":[{"cited_title":"First passage times for multiple par- ticles with reversible target-binding kinetics.The Journal of Chemical Physics , 147(13):134112, 2017","cited_arxiv_id":null,"evidence_quote":"Provides the analytical treatment of two independent particles with reversible target-binding kinetics that the non-interacting limit builds on."},{"cited_title":"Sequence heterogene- ity and the dynamics of molecular motors","cited_arxiv_id":null,"evidence_quote":"Supplies the kinetic Monte Carlo simulation method used for the general-regime results, including the non-monotonic curve."},{"cited_title":"Speed- selectivity paradox in the protein search for targets on dna: is it real or not? The Journal of Physical Chem- istry B, 117(42):12695–12701, 2013","cited_arxiv_id":null,"evidence_quote":"Gives the single-particle mean first-passage time formulas for a 1D random walk on an interval used in the non-interacting and strong-attraction limits."},{"cited_title":"Intermittent search strategies","cited_arxiv_id":null,"evidence_quote":"Supplies the intermittent-search concept used to explain why mixing 1D and 2D search modes can be faster than either alone."},{"cited_title":"First-passage times for random walks in bounded domains","cited_arxiv_id":null,"evidence_quote":"Provides the first-passage results for random walks in bounded domains used for the $L^2 \\ln L$ scaling when the target acts as a reflecting boundary."}],"review_version":1}