{"id":"99279d01-41f0-45b0-b864-4305915fef4f","arxiv_id":"2607.29599","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"In dilute quenched Poisson receptor fields, multivalent binding selectivity equals the mean receptor excess under bound particles; non-Poisson fluctuations add a Fano-factor or count-resolved correction.","lead":"Using multivalent particles, the paper shows that sharp 'superselective' binding to random receptor surfaces equals the average number of extra receptors sitting under each bound particle. That makes selectivity measurable from a single co-registered receptor–particle image rather than a noisy titration derivative.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified","rationale":"The paper's central claim is conditional on the dilute quenched Poisson regime, where the direct occupation term vanishes by construction. The derivation of Eq. 1 is an exact calculus identity, and for Poisson statistics R = n − n̄ is exact. The simulations provide independent confirmation via finite differences of the coverage. The paper explicitly discusses the limitation when particle crowding creates a density-dependent θ, and shows the deviation in Fig. 3, so the claim is not overextended. The single-image estimator is a practical corollary that requires sufficient occupied footprints, which the paper notes. There is no circular reasoning or hidden assumption; the weakest assumption is precisely the one the reader flagged, but it is stated and tested. Therefore the ACCEPT verdict stands.","tokens_in":14874,"tokens_out":10220,"duration_ms":106328,"concrete_test":"Run the same lattice-MC scan with a qualitatively different binding kernel, e.g., a threshold law θ(n) = 1_{n ≥ m} instead of the multivalent weight in Eq. S2, at the same z = 10^{-8}. If α from finite differences still matches ⟨n⟩_occ − n̄ across the Poisson scan, it confirms the identity's independence of binding details; any mismatch would indicate a hidden dependence of θ on n̄.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim α = ⟨n⟩_occ − n̄ is derived rigorously from the exact decomposition (Eq. 1) under the explicitly stated condition that the direct occupation term vanishes. The paper verifies the condition in Monte Carlo simulations across multiple receptor statistics and identifies the precise failure mode (particle crowding at high coverage) in Fig. 3, showing it enters through the direct term. The identity is independent of the binding law θ(n) as long as θ(n) is independent of n̄, making it robust. No internal inconsistency or missing justification was found.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper derives an exact decomposition of multivalent-binding selectivity α=∂lnΘ/∂ln n̄ into an occupied-footprint average of the receptor-count response R(n;n̄)=∂lnP(n;n̄)/∂ln n̄ plus a direct occupation term (Eq. 1). For quenched receptors and independent guest footprints the direct term vanishes. For Poisson receptor statistics R=n−n̄, so α equals the mean receptor excess beneath bound particles (Eq. 2), giving a single-density image-based estimator. For general count distributions with a linear count response the excess is rescaled by the Fano factor (Eq. 3); the paper introduces M(n̄) as the squared correlation between n and R, measuring how much of the Fisher information lies in the linear count mode. Lattice simulations verify the relations for Poisson, binomial, binary Cox, and interacting receptor ensembles, and off-lattice simulations show that guest-particle crowding appears through the direct occupation term.","tokens_in":15017,"tokens_out":9001,"duration_ms":97260,"significance":"The central identity is exact, parameter-free, and surprising: it links a macroscopic logarithmic derivative to a microscopic conditional mean. The paper therefore gives a fluctuation-response interpretation of superselectivity and a practical recipe for estimating α from co-registered images without differentiating a titration curve. The simulations are thorough: independent replicas, bootstrap intervals at the replica level, explicit protocols, and honest treatment of the regime where the direct occupation term does not vanish (Fig. 3, S1.E). The nonlinear-response diagnostic M(n̄) is a useful, falsifiable quantity. If the result is correct—and the derivation and checks support it—this will be a reference for the field.","major_comments":[],"minor_comments":[{"comment":"The phrase \"fraction of the density sensitivity captured by the count excess\" is imprecise: M is a squared correlation, i.e. the fraction of Fisher information carried by the linear projection, not a fraction of α itself. S1.C clarifies this, but the first mention should be reworded to avoid misreading.","section":"Beyond Poisson, Eq. (4)"},{"comment":"The condition for the direct term to vanish is better stated as \"independent footprints\" or \"negligible particle–particle correlations\" rather than \"dilute particles\". In the lattice model the direct term vanishes by construction even at finite coverage, while the off-lattice failure in S1.E and Fig. 3 is specifically guest-particle crowding; adjusting the wording would make the operating assumption precise.","section":"Poisson receptor fluctuations, before Eq. (2)"},{"comment":"At the phase-conversion points, both reconstructed derivatives are resolution-sensitive, as the authors state. The main-text Fig. 2(c) would benefit from a visible marker or shading of this narrow interval so that readers do not interpret the pointwise values there as reliable continuum derivatives.","section":"Interacting receptors and phase conversion, S2.D"}],"recommendation":"minor_revision","confidential_remarks":"No concerns beyond the minor wording and presentation issues in the report. The paper is technically sound, transparently identifies its own validity limits, and is suitable for publication after these local improvements."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe one thing to know: this paper is a solid theory contribution, not a breakthrough. It derives an exact decomposition of superselectivity and shows that in the dilute quenched Poisson limit, α equals the mean receptor excess under bound particles. Eq. 1 is essentially a logarithmic-derivative identity—you differentiate the coverage after averaging over the receptor-count distribution—but the value is in what the authors extract from it: an occupied-footprint interpretation, a Fano-rescaled formula, a linear-mode weight M(¯n), and a single-density image estimator.\n\nWhat is genuinely good: the paper does not oversell the mathematics. It states the condition under which the direct occupation term vanishes, tests it across Poisson, binomial, Cox, interacting-lattice, and off-lattice hard-disk receptor statistics, and then shows the failure mode—particle crowding at high coverage—entering through the direct term, exactly where you would expect it. The simulation protocols are detailed, the bootstrap intervals are computed over complete replicas rather than naive windows, and comparing finite-difference, Fano-rescaled, and count-resolved estimates side by side is useful and honest. The M(¯n) diagnostic is a nice tool: it tells you when the linear count mode is sufficient and when shape changes matter. No fitted parameters, no circular fitting, clean citation pattern including the author's own prior receptor-uniformity work where it is directly relevant.\n\nSoft spots, in proportion. The central identity is exact, so the physical assumptions carry the weight. The single-image estimator assumes Poisson statistics and enough occupied footprints; the paper mentions the sampling requirement but does not quantify the field-of-view or particle count needed for realistic imaging. The Cox and interacting-lattice examples are illustrative, and the near-phase-separation results are resolution-sensitive—the authors flag this, but those panels are less decisive than the Poisson cases. Calling Eq. 3 the \"Fano-rescaled\" result for a linear response is somewhat tautological: if R(n;n̄) is exactly linear, the formula follows by construction. But M(¯n) gives you a way to check that linearity, so the framework stands.\n\nThe stress-test note is right: I do not see a missing justification or a load-bearing flaw. This paper deserves a serious referee, not a desk rejection. It will be useful to experimental groups working on multivalent targeting and to theorists modeling receptor organization. I would cite it for the occupied-excess identity and the M diagnostic, and I would bring it to a reading group.","headline":"A sound, honest theory paper: the central Poisson identity is exact and the limits are clearly marked, though the novelty is interpretive and practical rather than conceptual.","tokens_in":15438,"tokens_out":2337,"would_cite":true,"duration_ms":28302,"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":"This paper shows that superselectivity in multivalent binding measures, in the dilute limit, the mean excess of receptors beneath bound particles — readable from a single co-registered image instead of a fitted titration curve.","keywords":["superselectivity","multivalent binding","receptor fluctuations","fluctuation-response relation","Poisson statistics","Fano factor","selectivity","adsorption"],"falsifier":"On a quenched, randomly grafted surface, measure selectivity two ways: from the slope of the coverage–density titration and from the single-image difference ⟨n⟩_occ − n̄ (Fano-rescaled when needed). If the two disagree beyond sampling error at low particle activity, the central identity fails. The paper's off-lattice simulations already locate one failure mode — at high coverage the titration slope falls below the image estimate — so the decisive experiment is to reduce coverage until the discrepancy disappears.","tokens_in":14814,"feed_emoji":"🎯","tokens_out":12561,"duration_ms":111396,"temperature":0.7,"pith_summary":"Multivalent particles — viruses, colloids, designed polymers — bind surfaces through many weak contacts, and superselectivity is their ability to switch from almost no binding to strong binding within a narrow range of receptor density. This paper asks what the standard measure of that sharpness, the logarithmic response α = ∂ln Θ / ∂ln n̄, actually measures at the microscopic level, and answers with an exact decomposition into two terms: how receptor-count states are reweighted as the mean density rises (averaged over bound particles), and how binding probability at a fixed receptor count changes directly along the scan. In the dilute limit on a fixed 'quenched' receptor landscape the second term vanishes, and for Poisson-distributed receptors α reduces to the mean number of excess receptors beneath bound particles — a quantity that can be read off a single co-registered receptor–particle image, no titration or numerical derivative required. Beyond Poisson statistics, the same excess is rescaled by the Fano factor when the count response is linear, and a new quantity M(n̄) says when nonlinear reshaping of the count distribution makes the simple excess insufficient. If correct, this turns superselectivity from a fitted curve parameter into an observable fluctuation property of the receptor landscape.","feed_headline":"Superselectivity equals the receptor excess beneath bound particles","feed_subtitle":"The new identity equates selectivity with mean receptor excess — measurable from a single image.","key_machinery":"The load-bearing object is the count-resolved response R(n;n̄) = ∂ln P(n;n̄)/∂ln n̄, which measures how the probability of a footprint containing n receptors changes as the mean density rises. Selectivity is the average of this response over occupied footprints, because binding preferentially samples receptor-rich environments. The exact decomposition α = ⟨R⟩_occ + ⟨∂ln θ/∂ln n̄⟩_occ separates receptor reweighting from direct occupation changes; for Poisson counts R = n − n̄, so α reduces to ⟨n⟩_occ − n̄, a difference of two means measurable from one image. The Fano factor rescales the excess for linear non-Poisson responses, and the linear-mode weight M(n̄) = Corr²[n,R] reports how much of","core_discovery":"The paper claims that in the dilute quenched regime, superselectivity is a fluctuation–response relation: selectivity α equals the occupied-footprint average of the count-resolved response R(n;n̄) = ∂ln P(n;n̄)/∂ln n̄, the direct occupation term of the exact decomposition vanishing. For Poisson receptor statistics this reduces to α = ⟨n⟩_occ − n̄: each bound particle sits, on average, on a footprint with more than one excess receptor. Beyond Poisson, a linear count response divides the same excess by the Fano factor; clustering or frozen heterogeneity requires the full count-resolved response, averaged over occupied footprints. Simulations confirm each relation, and the decomposition locates","pith_inferences":["The framework yields a practical classification rule the paper does not spell out: measure M(n̄) on a receptor-only density series, then trust the single-image estimator only where M ≈ 1; where it drops, reconstruct R(n;n̄) from two neighboring histograms and average it over occupied footprints.","A testable separation follows for surface engineering: two surfaces with identical mean receptor density but different fluctuation amplitude (different Fano factors) should show different selectivities, so titration curves alone cannot distinguish binding physics from receptor organization without fluctuation information.","In nonequilibrium or mobile-receptor settings, the vanishing of the direct occupation term is the natural place kinetics enters; the decomposition suggests the direct term carries exactly the time-dependent part of selectivity and could be used to phrase kinetic superselectivity in the same response language.","Because the estimator is a difference of two means, its precision is governed by the number of bound particles rather than curve smoothness — so with enough occupied footprints, selectivity could be mapped spatially across patterned or heterogeneous surfaces, a diagnostic the paper gestures at but does not develop."],"forward_implications":["A single co-registered receptor–particle image can replace multi-density titration for measuring selectivity in the dilute Poisson regime, removing derivative noise at low coverage.","Superselectivity gains a microscopic meaning: on Poisson landscapes each occupied footprint must contain, on average, more than one excess receptor for the binding curve to be sharp.","For finite-capacity or weakly correlated receptor fields, only the Fano factor is needed to correct the receptor-excess estimate; the full response is unnecessary.","On clustered or heterogeneous surfaces (Cox statistics, interacting receptors), the count excess and its Fano rescaling systematically underestimate the true response; receptor-only histograms at two neighboring densities supply the needed correction.","At high particle coverage, crowding adds a direct occupation response that no receptor-based measurement captures, delimiting the regime of validity of image-based estimators."],"fun_headline_variants":["Superselectivity = mean receptor excess under bound particles","Measure superselectivity from one large co-registered image","Receptor fluctuation response explains superselectivity","Selectivity equals local receptor excess in Poisson limit","Superselectivity as a fluctuation-response law"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The central premise is that the probability of a footprint with n receptors being bound does not depend on the global mean receptor density — true only for dilute particles on a fixed receptor landscape, and broken when adsorbed particles crowd and block one another, a breakdown the paper's own off-lattice simulations exhibit at high coverage (Supplemental S1.E, Fig. 3a).","fun_headline_variants_meta":{"raw":{"variants":["Superselectivity = mean receptor excess under bound particles","Measure superselectivity from one large co-registered image","Receptor fluctuation response explains superselectivity","Selectivity equals local receptor excess in Poisson limit","Superselectivity as a fluctuation-response law"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000228,"raw_usage":{"total_tokens":1272,"prompt_tokens":668,"completion_tokens":604,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":412,"completion_tokens_details":{"reasoning_tokens":542}},"tokens_in":412,"tokens_out":604,"duration_ms":6609,"temperature":1.0,"reasoning_tokens":542,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T03:47:22.321927+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a quenched, randomly grafted surface, measure selectivity two ways: from the slope of the coverage–density titration and from the single-image difference ⟨n⟩_occ − n̄ (Fano-rescaled when needed). If the two disagree beyond sampling error at low particle activity, the central identity fails. The paper's off-lattice simulations already locate one failure mode — at high coverage the titration slope falls below the image estimate — so the decisive experiment is to reduce coverage until the discrepancy disappears.","supporting_citations":[],"review_version":1}