{"id":"77c251a7-c8ef-4514-8d9c-c9547aed33ed","arxiv_id":"2501.02944","paper_version":7,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A minimal autocatalytic-cycle model derives bacterial growth laws and shows that reversible and irreversible antibiotic susceptibility regimes are generic, not specific to ribosome-targeting drugs.","lead":"This paper derives a minimal biophysical model in which antibiotic molecules slow bacterial growth by inhibiting an essential autocatalytic cycle, and shows that the model reproduces known growth laws as well as two distinct regimes of antibiotic susceptibility. It is worth reading because it suggests that these regimes, previously tied to ribosome-targeting drugs, are generic features of any inhibitor of an autocatalytic cycle.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The generic two-regime claim rests on assuming the antibiotic-targeted cycle is the limiting cycle; this assumption is asserted but not tested, and the full coupled-cycle system may behave differently.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: the derivation reduces the two-cycle autocatalytic network to a single limiting cycle. I agree with the reader's conditional verdict. The analytical core is coherent: the algebraic reductions in Eqs. 30, 34-39 check out, and the reversible/irreversible limits follow from the self-consistent Eq. 35. However, the central claim extends from an isolated single-cycle model to 'any inhibitors targeting an autocatalytic cycle,' and the bridge is precisely the single-limiting-cycle assumption. This is the most load-bearing point because it is the only place where the two-cycle coupling enters the derivation; the rest of the paper's analytic results are properties of one cycle. The empirical fits cannot resolve the issue because they are performed with the isolated model, the data and supplementary material are not accessible, and no code is shipped. A numerical solution of the full coupled system at the balanced-cycles condition would settle whether the generic regime classification survives when the min-function coupling is nontrivial. The secondary sign inconsistency between Eq. 12 and Eq. 54 is worth correcting but does not by itself change the verdict. My recommendation is therefore to keep the reader's CONDITIONAL verdict unchanged: the paper merits publication only after the scope of the generic claim is tested or explicitly qualified.","tokens_in":23074,"tokens_out":8738,"duration_ms":164136,"concrete_test":"Solve Eq. 20 numerically with the min function intact for a single antibiotic targeting B1, at parameters where the two cycles are balanced at aex=0 (kB1 B1u/Btot approximately kC1 C1u/Ctot). Vary aex and compute lambda(aex) and IC50(lambda0) (vary lambda0 via nutrient rates). Compare with the isolated-cycle prediction Eq. 49 and its limits (Eqs. 52, 54). If the reversible/irreversible regimes (monotone vs bistable response; sign of dIC50/dlambda0) are reproduced, the generic claim survives; if the coupled system shows plateaus, altered bistability, or shifted regime boundaries, the claim must be restricted to strictly limiting target cycles.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the reversible/irreversible regimes are generic for any inhibitor of an autocatalytic cycle is derived from the isolated, single-cycle self-consistent equation (Eq. 49). The isolation is justified only by the assertion in the section 'Modified model based on autocatalytic cycles': 'we assume the cycle targeted by the toxic agent becomes limiting... Consequently, we isolate the inhibited cycle and study its growth.' This is not a harmless technical step. In the full model (Eq. 20), the two cycles are coupled through min(kB1 B1u, kC1 C1u); if the drug-targeted cycle is not the limiting one, or if both branches of the min function are close to equality, the growth rate is set by the other cycle or by the switching between them, and Eq. 49 does not govern the dose-response curve. Appendix C.2 analyzes the second cycle only under the same 'B limiting' assumption and therefore does not cover this regime. The two-drug extension (Eq. 60) solves the coupled system, but no analogous single-drug solution or numerical exploration of the balanced-cycles regime is provided. Thus the generic statement is under-supported exactly where the model's coupling is nontrivial. A secondary internal inconsistency: Eq. 12 states Q(lambda)=1+Pinaex/lambda in the irreversible limit, whereas the Appendix derivation (Eq. 54) and Eq. 38 give Q(lambda)=1-Pinaex/lambda; the plus sign would imply growth enhancement by the inhibitor and appears to be a typo, but it signals that the irreversible-limit expressions in the main text need care.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a minimal biophysical model of bacteriostatic antibiotic action in which cell metabolism is represented as two coupled autocatalytic cycles (ribosome and RNA-polymerase production) interacting through a Leontief minimum function. The drug is assumed to bind to the autocatalyst of one cycle. The authors derive the two empirical growth laws rather than postulating them, recover the reversible and irreversible regimes of growth-dependent susceptibility previously found by Greulich et al., propose a risk proxy B1,b/B1,u, fit literature dose-response data for several antibiotics, and extend the framework to two-drug combinations and self-inhibiting waste. The central claims are that the two susceptibility regimes are generic for inhibitors of any autocatalytic cycle and that the growth laws follow from the model.","tokens_in":23368,"tokens_out":8878,"duration_ms":78519,"significance":"If the central claims hold, the model provides a mechanistic unification of bacterial growth laws and the empirical classification of antibiotics into reversible and irreversible binders, and it connects these to the hypothesis of autocatalytic cycles as the basic modules of cell metabolism. The derivation of the growth laws from the cycle equations (rather than imposing them phenomenologically) is a genuine methodological advance, and the model makes falsifiable predictions (universal IC50 curve, growth-rate bistability above a threshold concentration, antagonistic interactions for two drugs acting on separate cycles). The paper also ships explicit derivations in Appendices A-C, which I verified for internal consistency. The significance is, however, conditional on closing the gap between the isolated-cycle analysis and the full coupled system, because the genericity claim is the main novel message.","major_comments":[{"comment":"The genericity claim ('the two regimes ... should in fact be expected generically for any inhibitors targeting an autocatalytic cycle') is derived from the isolated-cycle equation (Eq. 49), which is obtained by assuming that the drug-targeted cycle is the limiting cycle ('we assume the cycle targeted by the toxic agent becomes limiting'). In the full coupled system (Eq. 20), the cycles interact through min(kB1 B1u, kC1 C1u); if the targeted cycle is not the one setting the min, or if the two branches are close to equality, the growth rate is set by the other cycle and Eq. 49 does not describe the dose-response. Appendix C.2 still assumes B is limiting, and the two-drug system (Eq. 60) is not used to check the balanced-cycles regime for a single drug. To support the central claim, the authors should either prove that any bacteriostatic inhibitor makes its target cycle limiting, add a numerical analysis of the full system showing the same two regimes for the balanced case, or explicitly restrict the genericity statement.","section":"Modified model based on autocatalytic cycles; Conclusion"},{"comment":"The irreversible limit is written inconsistently: main-text Eq. (12) gives Q(lambda) = 1 + P_in a_ex/lambda, whereas the derivation from Eq. (49) in Appendix A.5.2 gives Q(lambda) = 1 - P_in a_ex/lambda, which is also consistent with the root of Eq. (38), q = (1 + sqrt(1 - 4 P_in a_ex/lambda0))/2 approximately equal to 1 - P_in a_ex/lambda0. The plus sign would mean that increasing antibiotic concentration increases the fraction of active ribosomes, contradicting the model's purpose. This appears to be a typographical error, but it appears in a central display equation and must be corrected, with the surrounding text checked for further sign inconsistencies.","section":"Main text Eq. (12); Appendix A.5.2 Eq. (54)"},{"comment":"The statement that the model 'describes well' a large panel of antibiotics is not quantitatively supported. Each dose-response curve is fitted with four free parameters (Table 2) plus two globally hand-set values (KD = 1/50, kof = 5 h^-1), and no confidence intervals, goodness-of-fit measures, or model-comparison tests are reported. Without these, the reader cannot judge whether the observed concavities and the data collapse in Fig. 4b are meaningful evidence for the two-regime classification or merely a consequence of the fitting flexibility. A sensitivity analysis of the hand-set parameters is also needed.","section":"Experimental test of the model; Appendix B.2"}],"minor_comments":[{"comment":"The text reads 'where where r_b is the concentration' - the word 'where' is duplicated.","section":"Eq. (2)"},{"comment":"Several instances of 'k of f' should be 'k_of' (subscripts are corrupted in the text; e.g., Eq. (4) and surrounding text).","section":"Throughout"},{"comment":"In the last equation for dA/dt, the term '+k_of f B_1,u' should presumably read '+k_of f B_1,b', as in the simplified system Eq. (21); as written, the binding/unbinding balance is incorrect.","section":"Appendix A.5, Eq. (40)"},{"comment":"Reference [28] is an incomplete self-reference ('Supp. Mat. at ...; 2025'); provide a working link or a proper title for the supplementary material.","section":"References"},{"comment":"The text says 'we rescale the risk by P_in a_ex/(K_D P_out) to obtain a collapse of the experimental data in the reversible limit', but Fig. 3b shows a collapse of model curves, not of experimental data; please clarify the wording.","section":"Cell risk induced by the antibiotics"},{"comment":"The parameter m ('the number of limiting intermediate steps') is introduced rather quickly; please define it explicitly at first occurrence and state how it relates to the number of intermediate steps N.","section":"Appendix A.5.4"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious theoretical contribution with a solid core derivation, but the main novelty — the genericity of the two regimes — is broader than the analysis actually performed. The authors should either add a numerical study of the full coupled-cycle system or soften the concluding claim. The empirical fitting section needs error estimates and a sensitivity analysis. I do not see grounds for rejection, but the revision is substantial."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is a theory paper with a coherent analytical core, and the empirical comparisons are illustrative, not confirmatory. The authors derive the Scott-Hwa growth laws from a two-cycle autocatalytic model with a Leontief min coupling, and show the Greulich reversible/irreversible dose-response equations emerge as limits. That derivation is the real contribution. I checked the main reductions (Eqs 30, 34-39, 49, 52, 54) and they follow from the stated assumptions. The paper is honest about building on Roy-Pugatch and Greulich; the recovery of [13]'s equations is a consistency check, not a flaw.\n\nThe strongest claim—that the two susceptibility regimes are generic for any inhibitor of an autocatalytic cycle—is under-supported. It comes from isolating the inhibited cycle under the assertion that it becomes limiting. That is plausible for strong inhibitors, but the paper doesn't test the coupled system where both cycles are near co-limiting. Appendix C.2 only handles the B-limiting case. So 'generic' should be softened to 'generic within the single-limiting-cycle assumption.' This is a moderate soft spot, not a fatal one.\n\nThe empirical section is the weakest part. Fits use 4 free parameters per drug plus global hand-set KD=1/50 and kof=5 h^-1, with no error bars or goodness-of-fit. No data or code is shipped. That makes 'confirms the existence of two regimes' too strong; the data are consistent with the model, but I wouldn't call it confirmation. If the authors release the fitted data and scripts, that would help.\n\nThere is a definite sign inconsistency. Eq. 12 in the main text gives Q=1+Pinaex/lambda in the irreversible limit, which would mean growth enhancement by the inhibitor. Appendix A.5.2 (Eq. 54) and the derivation at Eq. 38 give Q=1-Pinaex/lambda. The plus sign is presumably a typo, but it signals the irreversible-limit expressions need careful proofreading.\n\nWho this is for: biophysicists working on growth laws and antibiotic susceptibility; also people building autocatalytic-cycle models. The risk proxy and the two-drug antagonism extension are useful ideas. The paper deserves a serious referee, and I'd accept it for review despite my reservations. The theory is worth engaging with; the authors just need to fix the sign, soften the generic claim, and be more careful in describing the empirical support.","headline":"Solid autocatalytic-cycle derivation of growth laws and Greulich regimes, but the 'generic two-regime' claim overreaches and the empirical support is thinner than advertised.","tokens_in":23983,"tokens_out":2411,"would_cite":true,"duration_ms":24685,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["92C40","92C37","37N25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that the two regimes of antibiotic susceptibility previously tied to ribosomes are generic to any inhibitor of an autocatalytic cycle, and derives the growth laws behind them from a minimal model rather than assuming them.","keywords":["bacterial growth","bacteriostatic antibiotics","autocatalytic cycles","growth laws","reversible and irreversible inhibition","growth-rate bistability","half-inhibitory concentration","antibiotic risk proxy"],"falsifier":"Find an antibiotic whose target is an essential autocatalytic cycle but whose half-inhibitory concentration, measured across growth media that change the inhibitor-free growth rate, cannot be brought onto the model's U-shaped IC50 curve (Eq. 15) with any choice of the model's rates; alternatively, observe a dose-response curve in the irreversible regime that is smooth with no branch jump near the predicted threshold concentration. Either observation would falsify the claim that the reversible/irreversible classification is generic.","tokens_in":22774,"feed_emoji":"💊","tokens_out":7135,"duration_ms":67730,"temperature":0.7,"pith_summary":"This paper proposes that bacteriostatic antibiotics slow bacterial growth by inhibiting one essential autocatalytic cycle in cell metabolism, modeled as two coupled cycles that make ribosomes and RNA polymerase. From that minimal picture, the authors derive the known bacterial growth laws instead of taking them as inputs, and they show that the reversible and irreversible regimes of growth-dependent susceptibility previously found only for ribosome-targeting drugs arise generically for any inhibitor of an autocatalytic cycle. The model also predicts a window of antibiotic concentration in which two growth rates coexist, a bistability that has been seen experimentally, and introduces a simple risk proxy: the ratio of drug-bound to free active enzymes. If the central claim is right, drug dose-response data for many antibiotic classes should collapse onto one generic curve relating half-inhibitory concentration to pre-exposure growth rate, with the drug's reversible or irreversible binding deciding which side of the curve it occupies.","feed_headline":"Two antibiotic response regimes are generic, not ribosome-specific","feed_subtitle":"A minimal autocatalytic-cycle model derives bacterial growth laws and predicts bistable growth for any targeted cycle.","key_machinery":"The machinery is a minimal network of two coupled autocatalytic cycles, one for ribosomes and one for RNA polymerase, where joint production steps obey a Leontief minimum rule: the rate of any step using two resources is set by the scarcer of the two. When the drug targets the currently limiting cycle, the model reduces to a single cycle with an arbitrary number of assembly steps, and the central object is the self-consistent equation for the growth rate lambda written through the fraction of active, unbound autocatalysts Q(lambda). Equating two expressions for the bound fraction B1,b/Btot gives Eq. (49), whose roots are the accessible growth rates; in the fast-assembly, long-lifetime limit it simplifies to a cubic that reproduces the earlier ribosome-specific equation, from which the reversible and irreversible limits are read off. The same self-consistent relation carries the two growth laws, the IC50 curve, the bistable window, and the risk proxy B1,b/B1,u, so the argument rides on this single equation.","core_discovery":"The paper's central claim is that antibiotic action does not need to be described ribosome by ribosome: any toxic agent that inhibits an essential autocatalytic cycle should produce one of two generic responses. In the reversible regime, where drug molecules leave the cell and unbind quickly, the growth rate declines smoothly with external drug concentration; in the irreversible regime, where drug accumulates and remains bound, a discontinuous transition to near-zero growth appears. Both behaviors follow from a single self-consistent equation for the growth rate, written through the fraction of unbound active autocatalysts. The derived growth laws, ribosome fraction increasing with growth rate in clean medium and decreasing when translation is inhibited, match the empirical relations that earlier work used as assumptions. The same equation yields a U-shaped dependence of the half-inhibitory concentration on the inhibitor-free growth rate, and measured values for several antibiotics, including one usually classified as bactericidal, are shown to collapse onto this curve. The authors conclude that the reversible/irreversible distinction and the growth-rate heterogeneity it produces should be expected generically for any inhibitor targeting an autocatalytic cycle.","pith_inferences":["Because the model is formulated in molecule numbers and uses only autocatalytic stoichiometry plus a minimum rule, the same reversible/irreversible dichotomy should appear in other autocatalytic systems exposed to poisons, such as ecological autocatalytic loops or economic production networks; the authors mention but do not develop this generality.","The constant ribosome density assumption ties cell volume to total mature ribosome count; if experiments show that ribosome density itself changes with drug concentration, the universal IC50 curve should bend or split, giving a testable boundary for the claim.","A stochastic version of this deterministic model, which the authors call for, would likely convert the coexistence window into a bimodal single-cell growth-rate distribution; measuring single-cell growth rates near the predicted threshold concentration could distinguish true bistability from population averaging.","The risk proxy suggests a quantitative ranking of antibiotics by early growth suppression independent of the bactericidal/bacteriostatic label, but testing it would require time-resolved measurements of bound target fractions, not just growth-rate dose-response curves."],"forward_implications":["Diverse bacteriostatic drugs, regardless of their molecular target, should fall into either the reversible or the irreversible class, with faster growth making the cell more susceptible in the reversible class and less susceptible in the irreversible class.","The known bacterial growth laws are consequences of autocatalytic-cycle balance rather than empirical constraints, and they should show slight curvature; departures from linearity in growth-law data are therefore expected, not noise.","Below a threshold external drug concentration, the model permits two coexisting growth rates, so growth-rate heterogeneity is an intrinsic property of the deterministic dynamics rather than a single-cell noise effect.","The ratio of drug-bound to free target molecules is a drug-class-independent risk proxy that rises sharply near IC50 and jumps discontinuously in the irreversible regime, making it a candidate measure for comparing antibiotic lethality.","Two drugs that target different coupled cycles act antagonistically: only the drug acting on the currently limiting cycle reduces growth, producing a dose-response surface that switches sharply between the two drugs."],"supporting_citations":[{"why":"Supplies the ribosome-targeting model with its reversible and irreversible regimes, the IC50 susceptibility curve, and the experimental dose-response data that this paper generalizes and refits.","marker":"[13]"},{"why":"Supplies the autocatalytic-cycle representation of cell metabolism and the minimum-function coupling that the model is built on.","marker":"[18]"},{"why":"Provides the empirical growth laws for ribosome fraction versus growth rate that the model derives from its autocatalytic cycles.","marker":"[17]"},{"why":"Defines the first growth law and the quantities such as translation capacity and ribosome fractions that the model reproduces.","marker":"[15]"},{"why":"Supports the link between bound ribosome fraction and drug cidality, motivating the paper's risk proxy.","marker":"[14]"},{"why":"Provides experimental data for antibiotic dose-response and growth conditions used to fit the model for several drugs.","marker":"[9]"},{"why":"Gives the existing model of combined antibiotic action that the two-drug extension compares against and extends to separate cycles.","marker":"[22]"}],"fun_headline_variants":["Generic antibiotic response from autocatalytic cycle model","Two regimes of antibiotic action, not just ribosome","Minimal model predicts universal antibiotic growth response","All antibiotics may share two response regimes","Autocatalytic cycle model explains antibiotic growth laws"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument holds only if the antibiotic slows growth by throttling one particular self-reproducing production loop that is the bottleneck, with ribosome density held fixed; if several loops limit growth together, the drug hits more than one loop, or ribosome density responds to the drug, the single-equation reduction and the generic two-regime conclusion do not follow.","fun_headline_variants_meta":{"raw":{"variants":["Generic antibiotic response from autocatalytic cycle model","Two regimes of antibiotic action, not just ribosome","Minimal model predicts universal antibiotic growth response","All antibiotics may share two response regimes","Autocatalytic cycle model explains antibiotic growth laws"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000519,"raw_usage":{"total_tokens":2480,"prompt_tokens":880,"completion_tokens":1600,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":496,"completion_tokens_details":{"reasoning_tokens":1542}},"tokens_in":496,"tokens_out":1600,"duration_ms":10391,"temperature":1.0,"reasoning_tokens":1542,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:00:19.446044+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find an antibiotic whose target is an essential autocatalytic cycle but whose half-inhibitory concentration, measured across growth media that change the inhibitor-free growth rate, cannot be brought onto the model's U-shaped IC50 curve (Eq. 15) with any choice of the model's rates; alternatively, observe a dose-response curve in the irreversible regime that is smooth with no branch jump near the predicted threshold concentration. Either observation would falsify the claim that the reversible/irreversible classification is generic.","supporting_citations":[{"cited_title":"Growth-dependent bacterial susceptibility to ribosome-targeting antibiotics","cited_arxiv_id":null,"evidence_quote":"Supplies the ribosome-targeting model with its reversible and irreversible regimes, the IC50 susceptibility curve, and the experimental dose-response data that this paper generalizes and refits."},{"cited_title":"A unifying autocatalytic network-based framework for bacterial growth laws","cited_arxiv_id":null,"evidence_quote":"Supplies the autocatalytic-cycle representation of cell metabolism and the minimum-function coupling that the model is built on."},{"cited_title":"Bacterial growth laws and their applications","cited_arxiv_id":null,"evidence_quote":"Defines the first growth law and the quantities such as translation capacity and ribosome fractions that the model reproduces."},{"cited_title":"Kinetics of drug–ribosome interactions defines the cidality of macrolide antibiotics","cited_arxiv_id":null,"evidence_quote":"Supports the link between bound ribosome fraction and drug cidality, motivating the paper's risk proxy."},{"cited_title":"Invariance of Initiation Mass and Predictability of Cell Size in Escherichia coli","cited_arxiv_id":null,"evidence_quote":"Provides experimental data for antibiotic dose-response and growth conditions used to fit the model for several drugs."},{"cited_title":"Minimal biophysical model of combined antibiotic action","cited_arxiv_id":null,"evidence_quote":"Gives the existing model of combined antibiotic action that the two-drug extension compares against and extends to separate cycles."}],"review_version":1}