{"id":"0d55bbf1-8586-481a-8d1b-6d5b30d2ac67","arxiv_id":"2505.19416","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Exact time-dependent formulas are derived for PDE models of sizer, timer, and adder cell size control, with a stability condition for the sizer steady state.","lead":"Cells use different rules to decide when to divide. This paper solves the equations for the three classic rules, sizer, timer, and adder, giving exact formulas for the size distribution and testing them against simulations and E. coli data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.2's asymptotic-stability criteria are never satisfied or verified by the paper's examples: A is infinite for v(s)=2s, and Δ<0 is never checked for any finite-A case.","rationale":"The characteristic-solution formulas (Theorems 3.1, 3.3–3.6) are broadly self-consistent: given α(t), the along-characteristic exponentials and delay equations are plausible, and reductions to renewal equations for α are standard. The reader's weakest assumption, that daughters are statistically independent, is a modeling simplification the authors explicitly acknowledge as future work in the Discussion; it restricts biological interpretation but does not invalidate the conditional mathematical statements. The more serious internal gap is the sizer stability theorem. Theorem 3.2 is presented as a proof of asymptotic stability, yet the quantities entering its sufficient condition are either infinite for the exponential growth used in the paper's main sizer example, or never evaluated for finite-A cases. Because the simulations throughout Section 3.2 use the strongly concentrated Beta(96,96) kernel, the theorem's discriminant condition may be impossible for those parameters. This matters because the reader's strongest-claim explicitly includes 'prove asymptotic stability of the sizer steady state.' If Theorem 3.2 has no verified instance, that portion of the central claim is unsupported rather than established. The proposed computation is a single, unambiguous check: evaluate A, P, Q, α, and B for a canonical parameter set and determine whether Δ<0 can hold. If it cannot, the authors should restrict the stability claim, prove a sharper sufficient condition, or demonstrate stability for at least one concrete biologically relevant case. This does not overturn the exact-solution contribution, so the appropriate verdict remains conditional pending that check.","tokens_in":32520,"tokens_out":15408,"duration_ms":140067,"concrete_test":"For the Figure 2 parameter set (v(s)=2s, sd=1, p=Beta(96,96)) and for a finite-A case (e.g., v(s)=1+s), compute A, P, Q, α, and B from equations (79), (80), and (99), then test whether A<∞, 2α>B, and Δ<0 hold. If A=∞ or Δ≥0 for every parameter set the paper presents, Theorem 3.2 has no verified instance and the stability claim should be downgraded to an unproved sufficient condition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The asymptotic-stability claim for the sizer mechanism is the least secure part of the central argument. Theorem 3.2 defines A as the maximum age at which a cell starting from any sb in [0,1] reaches s=1, with s(a;sb) solving ds/da=v(s,a). For the exponential growth v(s)=2s used in Section 3.2.3 and Figure 2, s(a;sb)=sb e^{2a}, so the first-passage age is (1/2)log(1/sb), which is unbounded as sb→0. Hence A=∞, the integral defining Q in (80) is not finite, and the condition Δ<0 is not a well-defined finite inequality. For the linear case v(s)=v0+v1s with v0>0, A is finite, but the paper never evaluates the hypotheses 2α>B and Δ<0 for any concrete parameter set. With the Beta(96,96) daughter kernel used throughout the simulations, p is strongly concentrated near s=1/2, so P in (80) is large and 1−4AP may well be negative; in that case Δ = AQ − (2α−B)(1−4AP) is necessarily positive, making the theorem inapplicable. Thus the claimed rigorous stability proof has no demonstrated instance. At best the theorem is an uninstantiated sufficient condition, and it does not support the Discussion's statement that stability is rigorously established under biologically realistic assumptions.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a first-order PDE model for the joint cell size-age distribution under three size-control strategies: sizer, timer, and adder, encoded through distinct boundary conditions. The authors derive characteristic-based exact solution representations for each mechanism (Theorems 3.1, 3.3, 3.5, 3.6), state a sufficient condition for asymptotic stability of the sizer steady state (Theorem 3.2), rederive the Collins-Richmond formula, and support the analytical results with individual-cell-based stochastic simulations and a fit to E. coli size distributions. The central mathematical claims are the exact time-dependent solutions, the closed-form marginal steady-state distributions for linear and logistic growth under sizer control, and the stability theorem for the sizer mechanism.","tokens_in":32900,"tokens_out":16040,"duration_ms":138441,"significance":"If the derivations are correct after fixing the issues below, the paper would be a useful contribution to the quantitative cell-size-control literature: it unifies the three classical mechanisms in one PDE framework, provides explicit characteristic representations and, for the sizer, closed-form marginals with verified normalization, and it transparently rederives a classical experimental formula. The stochastic simulations corroborating the analytic formulas and the application to E. coli data are also valuable. The main strengths are the self-contained derivations, the explicit parameter dependence of the steady-state distributions, and the fact that the analytic predictions are directly testable by simulation. However, the stability theorem and some theorem statements currently overstate what is proven, and these points must be corrected before the paper can be accepted.","major_comments":[{"comment":"The stability theorem's hypotheses are never verified for any concrete parameter set, and for the exponential growth v(s)=2s used in Section 3.2.3 and Figure 2, the quantity A defined as the maximum first-passage time to s=1 is infinite (s(a;0)=0 for all a, and the hitting time for sb>0 diverges as sb→0). Consequently, the bounds involving sqrt(A) in Eqs. (83)-(84) are not meaningful and Delta is not a well-defined finite expression. For the linear case v(s)=v0+v1s with v0>0, A is finite, but the paper does not evaluate the conditions 2α>B and Delta<0 for any parameter set; with the Beta(96,96) kernel used throughout and e.g. v0=1, v1=0.1, one has 4AP>1, so 1-4AP<0 and Delta>0, making the theorem inapplicable. The Discussion's statement that stability is \"rigorously established under biologically realistic assumptions\" (Section 4, paragraph 2) is therefore unsupported. The authors should either prove the conditions for a concrete biologically realistic parameter set or substantially soften the stability claim and state it only as a sufficient condition of unverified applicability.","section":"Section 3.2.4, Theorem 3.4"},{"comment":"The Discussion (Section 4, paragraph 3) claims that for the sizer, \"the size distribution at steady state is largely determined by p(s), and is independent of the detailed form of the cell growth rate v(s,a).\" This is contradicted by the paper's own Theorem 3.3 and Figure 3d: for v(s)=v0+v1s with both parameters nonzero, the steady-state distribution (100) depends on v(s) explicitly through v(s) and through α, and Figure 3d shows different distributions for different (v0,v1). Independence holds only in the special cases v1=0 (Eq. 107) and v0=0 (Eq. 109). The Discussion should be rephrased to state the precise parameter regimes in which the distribution is growth-rate-independent.","section":"Section 3.2.4, Theorem 3.4"},{"comment":"The boundary condition assumes that each mother produces two statistically independent daughters with a size distribution p(s,s') that does not depend on the mother's age or growth history. This assumption is load-bearing: the exact solution representations (Theorems 3.1, 3.5, 3.6) and the stability theorem (Theorem 3.2) all rely on the factorization of the birth kernel. The manuscript acknowledges this limitation in the Discussion, but it should be stated more prominently in Section 2 and the E. coli fit should not be interpreted as validating the independence assumption, since the fit only adjusts p(s,s') and Δs. This is a caveat, not a rejection, but it materially affects the biological interpretation of the derived distributions.","section":"Section 2.2, Eq. (22); Section 4, last paragraph"}],"minor_comments":[{"comment":"The soft-control timer division probability is written as ϕ(a)=1/(sqrt(2π)σ) e^{-(s-T)^2/(2σ²)}; the exponent should be (a-T)^2, consistent with the argument ϕ(a) and with Eq. (137).","section":"Section 3.3.3, Eq. (136)"},{"comment":"In Figure 3(b), the solid line for constant growth v(s)=v0 should reference Eq. (107), not Eq. (109); in Figure 3(d), the theory line should reference Eq. (100), not Eq. (109).","section":"Figure 3 captions"},{"comment":"The text says \"we adjusted the parameters c0, c1, and Δs\" and \"listed in (1)\"; it should say \"r, ā, b̄, and Δs\" and \"listed in Table 1.\"","section":"Section 3.4.2, Table 1"},{"comment":"Numerous typos need a careful proofreading pass: \"esquation\" in Section 2.1, \"satifies\" in Lemma 3.1, \"asymtotically\" in Theorem 3.2, \"sort control\" in Figures 5 and 7 captions, \"ell size\" in Section 3.3.2, \"Chepman-Kolmogorov\" in the Introduction, and reference errors such as \"J Asut Math Soc.\"","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's main novelty lies in the characteristic-based exact representations and the closed-form marginals for the sizer, but the stability theorem currently lacks any verified instance and is presented as stronger than it is. The E. coli fit is illustrative rather than statistically rigorous; if the journal expects quantitative model validation, the authors should add a goodness-of-fit measure. The paper may be better suited to a mathematical-biology venue than to a general biology audience, given the emphasis on PDE techniques."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick read: the exact time-dependent solutions for the sizer, timer, and adder PDEs are the real meat here, and they look like legitimate extensions of the steady-state work by Miotto, Jia, and Xia. The method-of-characteristics derivation in Theorem 3.1 is coherent, the linear and logistic special cases are useful, and the simulation/theory agreement in the sizer and timer sections is mostly convincing. You should know that before sending this anywhere, the stability theorem needs work.\n\nThe stability theorem is the weakest part. For the exponential growth v(s)=2s actually used in Fig. 2, the first-passage time to s=1 from sb=0 is infinite, so A=∞ and the quantities B, Q, and Δ in Theorem 3.2 are not well-defined. The paper never evaluates 2α>B and Δ<0 for any concrete parameter set with finite A. With a Beta(96,96) daughter kernel, P is large enough that 1−4AP is arguably negative, which would force Δ>0 and make the theorem inapplicable. So Theorem 3.2 is at best an uninstantiated sufficient condition, and the Discussion's claim of \"rigorously established stability under biologically realistic assumptions\" outruns the evidence.\n\nThere are also typos in key displayed formulas. Theorem 3.4 states v(s)=v1(1−s/v0), missing the factor s used in Eq. (110); Eq. (136) should have (a−T)^2, not (s−T)^2; and Theorem 3.4(3) gives 2α/(s v(s)) where the derivation gives 2α/v(s). These are fixable, but they sit in the main results.\n\nThe E. coli fit in Fig. 9 is a parameter calibration, not an independent validation: shape parameters r, a_bar, b_bar, and Δs are adjusted per temperature, with no error bars or goodness-of-fit. That's fine as an illustration, but it shouldn't be read as strong evidence for the adder model.\n\nThe independent-daughter boundary condition is a real assumption, but the paper flags it in the Discussion; within the model it's standard. It limits biological scope, not a hidden flaw.\n\nBottom line: this is a useful methods paper for the cell-size modeling community, but not ready as is. If the authors fix the stability theorem (make A finite, check the conditions for at least one realistic case, or downgrade the claim), correct the typos, and temper the Discussion, I'd be happy to see it in print. It deserves a serious referee rather than a desk reject, and I'd want the referee to push on Theorem 3.2.","headline":"Exact characteristic solutions for the three size-control PDEs are a genuine contribution, but the advertised stability theorem is never instantiated and the manuscript has enough typos in key formulas to need a serious revision before I'd trust it.","tokens_in":33350,"tokens_out":4854,"would_cite":false,"duration_ms":48226,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["92C37","35F15","35C05"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper derives exact time-dependent and steady-state solutions for sizer, timer, and adder cell-size control models, with explicit conditions for stability of the sizer steady state.","keywords":["cell size control","sizer mechanism","timer mechanism","adder mechanism","method of characteristics","size homeostasis","stochastic simulation","steady-state distribution"],"falsifier":"Measure mother–daughter size pairs in a population controlled by a sizer: if the two daughters of one mother show correlations beyond the fixed kernel $p(s,s')$, or if the observed homeostatic marginal distribution does not match $\\tilde f(s)=2\\alpha/v(s)\\int_0^s p(u)K_0(\\alpha,s,u)\\,du$ with $\\alpha$ from $2\\int_0^1 p(u)K_0(\\alpha,1,u)\\,du=1$, the renewal boundary condition that carries the whole derivation is empirically false.","tokens_in":32330,"feed_emoji":"🧫","tokens_out":7020,"duration_ms":60916,"temperature":0.7,"pith_summary":"The authors develop a single first-order partial differential equation model in which the three classic cell-size control strategies—sizer, timer, and adder—enter only through different boundary conditions. Using the method of characteristics, they derive exact time-dependent formulas for the cell size distribution under each mechanism, closed-form steady-state distributions for linear and logistic growth laws, and a stability criterion for the sizer steady state. The formulas reveal that the daughter-size inheritance rule, rather than the growth rate alone, largely determines the steady-state distribution. This matters because the equations turn cell-size homeostasis into quantitative, testable predictions, and the model reproduces experimental E. coli size distributions.","feed_headline":"Cell-size control reduced to exact distribution formulas","feed_subtitle":"Cell-size homeostasis follows from division rules, not growth rates alone.","key_machinery":"The load-bearing object is the linear transport equation $\\partial f/\\partial t+\\partial f/\\partial a+\\partial(v f)/\\partial s=-(\\alpha+\\phi)f$, where $f$ is the normalized density of cells of size $s$ and cycle age $a$, $v$ is the growth rate, $\\phi$ the division rate, and $\\alpha$ the population growth rate. Each control mechanism is enforced only through the boundary condition at birth ($a=0$) and through the singular division rate, and the solutions are built from the characteristic first integrals $t-a$ and $u(s,a)$, with the auxiliary functions $\\gamma$, $\\psi$, and the kernels $K_0,K_1$ encoding how initial size is transported to later ages.","core_discovery":"The central claim is that for a proliferating population with growth law $v(s,a)$ and daughter-size kernel $p(s,s')$, the full time-dependent cell size distribution can be written explicitly. For the sizer, the distribution along a characteristic is the boundary history $2p(\\gamma_0(s,a))\\alpha(t-a)$ times an exponential decay factor; for the timer it is the analogous renewal integral against $f(s',T,t-a)$; for the adder it is given in the birth-added-size coordinates $(s,\\varsigma)$ by a similar characteristic formula. At steady state these reduce to closed forms such as $\\tilde f(s)=2\\alpha/v(s)\\int_0^s p(u)K_0(\\alpha,s,u)\\,du$ for linear growth, with the population growth rate $\\alpha$ fixed by $2\\int_0^1 p(u)K_0(\\alpha,1,u)\\,du=1$. The paper further claims that the sizer steady state is asymptotically stable whenever $2\\alpha>B$ and $AQ-(2\\alpha-B)(1-4AP)<0$. These results give direct formulas connecting observable size statistics to the underlying division rule.","pith_inferences":["If the characteristic construction carries over to mixed strategies (division rate a sum of delta peaks), the same formalism would give exact solutions for hybrid sizer–timer controls without new mathematics.","The steady-state formulas suggest a data-inversion protocol: from measured birth and division size distributions one could reconstruct the inheritance kernel $p(u)$ nonparametrically, giving a direct test of which control mechanism a population uses.","The stability criterion $2\\alpha>B$ is checkable from lineage data if $\\alpha$ is estimated from population growth and $B$ is bounded by the age variation of the growth rate; this would separate stable from unstable sizer regimes empirically.","The paper's own discussion notes that sibling correlations are ignored; incorporating them would replace the scalar birth boundary condition with a joint two-daughter kernel, which would likely alter both the exact formulas and the stability threshold."],"forward_implications":["For the sizer with linear growth $v=v_0+v_1s$, the steady-state distribution depends on the ratio $v_1/\\alpha$ and the inheritance kernel $p(u)$, and the absolute growth rate drops out in the purely linear ($v_1=0$) and purely exponential ($v_0=0$) limits.","A pure timer with exponential growth ($v_0=0$) cannot maintain a stable positive cell size; under linear growth $v_0+v_1s$ a stable positive size requires $1<e^{v_1T}<2$ and $v_0>0$.","The adder model converges to a steady-state distribution that is insensitive to the growth rate, and with a truncated beta inheritance kernel it fits the measured E. coli size distributions at different temperatures, with $\\Delta s$ as the main temperature-dependent parameter.","Soft probabilistic division yields smoother steady-state distributions with tails beyond the hard threshold, while hard sizer and timer controls give strict cutoffs.","The exact characteristic formulas provide a direct iterative scheme to compute the time-dependent distribution from an arbitrary initial condition and the boundary history."],"supporting_citations":[{"why":"Supplies the Collins-Richmond formula rederived here, connecting homeostatic size distributions to the growth rate.","marker":"[10]"},{"why":"Provides the E. coli mother-machine size data used to fit the adder model.","marker":"[48]"},{"why":"Introduced the unified adder-timer PDE that the adder formulation extends.","marker":"[59]"},{"why":"Presents a size-dependent division strategy whose equation aligns with the sizer marginal model.","marker":"[32]"},{"why":"Models asymmetric division by a growth-fragmentation equation, the prior framework the boundary-condition approach builds on.","marker":"[60]"},{"why":"Provides a classical stability analysis of the cell size distribution that the sizer stability theorem generalizes.","marker":"[11]"},{"why":"Gives analytic results and parameter inference for lineage cell-size distributions across generations, motivating the renewal formulation.","marker":"[21]"}],"fun_headline_variants":["Exact formulas link cell size to division rules","Sizer, timer, adder: cell size distribution solved","Cell size distributions from boundary conditions exactly","Steady-state cell size from division kernels in closed form"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The boundary condition at birth assumes each mother produces two statistically independent daughters with sizes drawn from a fixed probability rule that does not depend on mother age or growth history; if daughter sizes are correlated or the rule varies with age, the exact solutions and stability theorem no longer hold.","fun_headline_variants_meta":{"raw":{"variants":["Exact formulas link cell size to division rules","Sizer, timer, adder: cell size distribution solved","Cell size distributions from boundary conditions exactly","Steady-state cell size from division kernels in closed form"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000237,"raw_usage":{"total_tokens":1475,"prompt_tokens":879,"completion_tokens":596,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":495,"completion_tokens_details":{"reasoning_tokens":535}},"tokens_in":495,"tokens_out":596,"duration_ms":6537,"temperature":1.0,"reasoning_tokens":535,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:14:17.800423+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure mother–daughter size pairs in a population controlled by a sizer: if the two daughters of one mother show correlations beyond the fixed kernel $p(s,s')$, or if the observed homeostatic marginal distribution does not match $\\tilde f(s)=2\\alpha/v(s)\\int_0^s p(u)K_0(\\alpha,s,u)\\,du$ with $\\alpha$ from $2\\int_0^1 p(u)K_0(\\alpha,1,u)\\,du=1$, the renewal boundary condition that carries the whole derivation is empirically false.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Collins-Richmond formula rederived here, connecting homeostatic size distributions to the growth rate."},{"cited_title":"Tanouchi, A","cited_arxiv_id":null,"evidence_quote":"Provides the E. coli mother-machine size data used to fit the adder model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced the unified adder-timer PDE that the adder formulation extends."},{"cited_title":"Miotto, S","cited_arxiv_id":null,"evidence_quote":"Presents a size-dependent division strategy whose equation aligns with the sizer marginal model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Models asymmetric division by a growth-fragmentation equation, the prior framework the boundary-condition approach builds on."},{"cited_title":"Diekmann, H","cited_arxiv_id":null,"evidence_quote":"Provides a classical stability analysis of the cell size distribution that the sizer stability theorem generalizes."}],"review_version":1}