{"id":"b764c010-5745-41a1-a70c-04e316bc7988","arxiv_id":"1908.06671","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Proteins that deposit marks they are repelled from can move directionally along DNA and dive into collapsed chromatin, speeding up target search.","lead":"This paper proposes a new mechanism, called chromophoresis, by which repair proteins could find their targets on DNA by depositing chemical marks and then moving away from those marks. If real, the same mechanism could let repair proteins penetrate dense, collapsed regions of the genome where DNA damage often hides.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central speedup claim rests on coverage per binding event, but no target is ever placed in the simulations and mean search time is never measured.","rationale":"The paper is a coherent computational study with a plausible Kramers-rate derivation for the 1D running state and a genuinely interesting qualitative observation in the collapsed-globule simulations: searchers can create marked trails, locally disrupt bridging-induced collapse, and dive into the globule. These parts of the argument are defensible. The load-bearing weakness I identify is different from the reader's: the reader focused on whether the required kon rate is physiologically achievable. That is a real concern for the biological extrapolation to repair proteins, but it is secondary to the paper's own physical claim of “fast target search”, because the manuscript never places a target in the simulations and never measures a search time. Coverage per binding event is a reasonable proxy, but it does not by itself establish the speed of finding a specific site, since event durations and off-fibre excursions enter the total search time. The conclusion that there is a “proved” optimal evaporation rate is stronger than the evidence supports. These gaps do not invalidate the mechanism, but they mean the headline claim should be read as a conditional proposal rather than a demonstrated speedup. The reader's verdict of CONDITIONAL already captures this level of caution, so I recommend no change to the verdict. A first-passage-time simulation would directly settle the issue and, if successful, would strengthen the paper substantially.","tokens_in":8249,"tokens_out":10791,"duration_ms":129136,"concrete_test":"Repeat the collapsed-globule simulations with one target bead positioned at a buried locus and N chromophoretic searchers, for koff values around the kc identified in Fig. 5A. Measure the mean first-passage time (or inverse hitting rate) to the target, including time spent off the fibre, and compare it with (i) the same system with kon = 0, i.e. no mark deposition and no chromophoresis, and (ii) a swollen-fibre passive search. If the minimum mean first-passage time does not occur near kc, or if the chromophoretic first-passage time is not lower than the passive control in the same geometry, then the claims of an optimal evaporation rate and of a fast target-search enhancement are not supported by the current data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract claims a “striking enhancement of the efficiency of target search” and the Conclusions state “we proved the existence of an optimal evaporation rate of epigenetic marks for which the exploration of the fibre is fastest.” The quantitative support for this in the collapsed-globule setting is Fig. 5A, which reports the average fraction of fibre beads visited per binding/diving event as a function of koff. Coverage per dive is not the same as target-search time: a dive has a duration, and search includes detachment, 3D diffusion, and rebinding. Indeed, Fig. 5B shows a change in the time spent on the fibre at the same kc, so the optimum in coverage could be offset by unfavourable event durations or rebinding statistics. The 1D section likewise argues speedup only verbally, from enlarged distance covered per binding, without reporting a first-passage time or a direct comparison with a symmetric random walk. No target bead is ever placed in the simulations, and no passive-facilitated-diffusion control is run in the same collapsed geometry. The claim that chromophoresis is “far faster” than passive search is therefore an extrapolation from a coverage proxy, not a measured search rate. The word “proved” also overstates what a nonmonotonic curve from 10–20 simulations can establish.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a nonequilibrium 'chromophoresis' mechanism for protein target search on chromatin: a DNA-binding protein deposits an epigenetic mark on nearby beads and is repelled from the mark it deposits, producing directed motion. In a 1D model the authors use Kramers theory to compute escape, hopping, and run-length rates, and show that sufficiently fast mark deposition kon > D1D/sigma^2 produces a unidirectional running state. In 3D Langevin simulations they report that multiple chromophoretic proteins spread along the fibre with suppressed density fluctuations, and that on a collapsed chromatin globule the proteins can locally unravel the globule and 'dive' into its core. The central quantitative claim is that there is an optimal mark-removal rate koff at which the fraction of the fibre visited per binding event is maximal, which the Conclusions phrase as having 'proved the existence of an optimal evaporation rate of epigenetic marks for which the exploration of the fibre is fastest.' The paper also speculates on relevance to PARylation and DNA repair.","tokens_in":8512,"tokens_out":4360,"duration_ms":46876,"significance":"If the central claims hold, chromophoresis would be a genuinely new and biologically suggestive search strategy, distinct from facilitated diffusion and applicable to collapsed chromatin where passive searches fail. The strengths of the paper include an analytic 1D Kramers derivation with an explicitly stated potential and no fitted target-search parameters, a clear separation between the 1D mechanistic argument and the 3D simulation phenomenology, and falsifiable predictions (hyperuniform spreading, an optimal koff, and globule diving). The proposed link to PARylation is speculative but appropriately framed. The main gap is that the headline measure of search efficiency is a visited-fraction proxy, not a measured search time to a target, and some collective claims are asserted without the standard quantitative diagnostics.","major_comments":[{"comment":"The central claim that chromophoresis yields an optimal mark-removal rate for fastest target search is not directly supported by the data. Fig. 5A measures only the average fraction of fibre beads visited per binding/diving event, while Fig. 5B shows that the residence time on the fibre also changes at the same kc; coverage per event is not equivalent to a search rate, because event duration, detachment, and 3D rebinding all contribute to the mean search time. No target bead is placed in the simulations and no first-passage time to a target is reported, so the Conclusions sentence 'we proved the existence of an optimal evaporation rate... for which the exploration of the fibre is fastest' overstates what a nonmonotonic coverage curve from 10-20 simulations can establish. I recommend adding direct measurements of mean search time to a randomly placed target (or at least a rate combining coverage and event duration) and softening the 'proved' language.","section":"3D Model, Fig. 5, Conclusions"},{"comment":"The hyperuniform-spreading claim is asserted but not quantitatively demonstrated. The text states that trail-mediated exclusion leads to 'hyperuniform spreading along the substrate' and suppressed 1D density fluctuations, yet Fig. 3 plots only the average number of bound proteins and an inset pair-correlation function; there is no measurement of the structure factor at small wavevectors or of the variance of particle number in intervals of increasing length, which are the standard diagnostics for hyperuniformity, and no error bars are shown. The claim that multiple proteins 'spread out along the fibre with suppressed 1D density fluctuations' therefore needs either a direct hyperuniformity diagnostic or a more modest formulation.","section":"Collective behaviours, Fig. 3 and following text"},{"comment":"The speedup claim is argued from the run length lrun ~ (3/2) exp(epsilon/kBT) rather than from a target-search observable. Enlarging the distance covered per binding event is not by itself a proof of faster search, since a directed run has a finite duration, terminates in detachment, and is followed by 3D diffusion and rebinding; in a finite fibre with multiple particles, trails may also introduce temporal correlations. I ask for a direct comparison of mean first-passage time to a target for the chromophoretic model versus a symmetric random-walk (facilitated-diffusion) control in the same geometry, or an analytic calculation of the mean search time from the 1D rates.","section":"1D approximation"},{"comment":"The biological feasibility of the running state rests on the condition kon > D1D/sigma^2 ~ 1 s^-1, which the authors acknowledge is 'compatible, albeit slightly faster' than typical modification rates. Since unidirectional motion and the diving phenomenon both depend on this threshold, the in vivo relevance claim is sensitive to this assumption; the manuscript should either provide more quantitative support for achievable kon values for relevant marks (e.g., PARylation) or explicitly frame the mechanism as requiring a kinetic regime whose biological occurrence remains to be demonstrated.","section":"1D approximation, parameter estimate"}],"minor_comments":[{"comment":"The abstract contains a typo: 'undirectional motion' should read 'unidirectional motion.'","section":"Abstract"},{"comment":"In the caption, 'fraction of fibre visitided per dive' should read 'fraction of fibre visited per dive.'","section":"Fig. 5 caption"},{"comment":"The text gives lrun = B/2C e^(epsilon/kBT) with B and C defined only by approximate numerical prefactors; consider writing the dimensionless expression explicitly and numbering the equations, as several rates and inequalities are currently referenced only verbally.","section":"1D approximation"},{"comment":"The inset showing the two-point correlation function has no labeled axes, and the direction of motion and normalization are not specified; please clarify.","section":"Fig. 3"},{"comment":"The manuscript uses both 'Kramer's' and 'Kramers' theory; please standardize to 'Kramers.'","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper's framing and topic are well suited to the journal, and the analytic 1D derivation is a solid core. My recommendation is driven entirely by the gap between the visited-fraction proxy and the claimed search-time optimum; a revised version that measures first-passage times to an explicit target and adds the missing hyperuniformity diagnostics would make the central claims defensible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a read. The chromophoresis idea is the real deal: a protein that deposits marks and is repelled by them gets a local asymmetry that rectifies its motion on a 1D substrate. The Kramers calculation is coherent and the run-length estimate follows naturally. The collective effects—collision-induced dissociation, trail-mediated exclusion, hyperuniform spreading—are plausible and supported by the correlation function data. And the diving into a collapsed globule is a genuinely new prediction; that alone should interest anyone who works on repair in heterochromatin.\n\nThe soft spots are real, though not fatal. The central claim of 'fast target search' is never directly tested. No target is placed in the simulations, no first-passage time is measured, and the support in Fig. 5 is fraction of fiber covered per binding event. That is a proxy, and the stress-test note has it right: if the residence time changes with koff (as Fig. 5B shows), the coverage optimum may not be the search-time optimum. The word 'proved' in the Conclusions overstates what a nonmonotonic curve from 10–20 simulations can establish. The paper should either run explicit target-search simulations or soften the language.\n\nA second, minor concern: the requirement kon > D1D/sigma^2, which the paper estimates as about 1 s^-1, is at the edge of measured PTM rates. The text acknowledges this, which is honest, but it means the running state in vivo is not a given. This is a matter of biological applicability, not internal consistency.\n\nThe citation pattern is fine; the PARylation connection is plausible and well-referenced. No code or data are deposited, which is a shame for a simulation paper, but the description is detailed enough to reproduce.\n\nOverall: this is a solid physics Letter with a real conceptual contribution. It deserves a serious referee. My recommendation would be to send it out, and to ask for either direct search-time measurements in the collapsed geometry or a clear statement that the speedup claim is based on a coverage proxy. I would cite it for the chromophoresis mechanism, not for the quantitative search-time enhancement.","headline":"A novel negative-feedback mechanism for protein search on chromatin, with a clean 1D analytic core, but the speedup claim needs direct search-time simulations before it is proven.","tokens_in":9007,"tokens_out":2367,"would_cite":true,"duration_ms":23756,"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":"The paper proposes chromophoresis, a nonequilibrium search mechanism in which repair proteins deposit a chemical mark and are repelled by their own mark, letting them run along DNA and dive into collapsed chromatin to reach buried lesions.","keywords":["chromophoresis","epigenetic marks","target search","facilitated diffusion","chromatin","DNA repair","nonequilibrium dynamics","PARylation"],"falsifier":"Track a single fluorescently labeled chromatin-binding protein as it deposits marks on a collapsed, bridging-protein-folded chromatin globule in vitro. The central claim predicts persistent unidirectional runs and repeated inward-diving events once the mark-deposition rate exceeds about 1 s$^{-1}$; observing only diffusive sliding and surface sticking at any deposition rate would disprove the chromophoretic mechanism.","tokens_in":8043,"feed_emoji":"🧬","tokens_out":10477,"duration_ms":91048,"temperature":0.7,"pith_summary":"The paper proposes that repair proteins can find DNA lesions far faster than passive diffusion allows by depositing chemical marks on chromatin and then being repelled from those same marks. Because each protein creates its own repulsive trail, it moves persistently in one direction along the DNA, a state the authors call running, and on a collapsed chromosome the trail carves a local opening that lets the protein dive into the interior. Using a 1D analytic model and 3D Brownian dynamics simulations, the authors show that this nonequilibrium mechanism, chromophoresis, spreads searchers evenly along the genome and, at an optimal mark-removal rate, maximizes the fraction of chromatin visited per binding event. If real repair proteins such as PARP operate this way, the mechanism would explain how lesions buried in heterochromatin become accessible within minutes.","feed_headline":"Repair proteins that flee their own marks could find buried DNA","feed_subtitle":"Simulations show how self-deposited mark trails let repair proteins dive into collapsed chromatin and reach lesions fast.","key_machinery":"The load-bearing object is the negative-feedback chromophoretic cycle. A bound protein deposits a mark on a neighboring chromatin bead at rate $k_{\\rm on}$; the mark abrogates the protein's attraction to that bead, creating an asymmetric potential in which the protein slides downhill on the unmarked side at rate $q_+ \\sim \\epsilon$ and hops backward at the much smaller rate $q_- \\sim \\epsilon e^{-\\epsilon/k_B T}$. Iterating this cycle converts symmetric sliding into a run with length $l_{\\rm run} \\simeq (B/2C)e^{\\epsilon/k_B T}$, and the trail of marks left behind repels other searchers, producing hyperuniform spreading. On a collapsed globule, the actively maintained gradient of unmarked beads toward the interior drives the diving motion, and the optimal evaporation rate $k_{\\rm off}$ sets a cusp in the visited fraction of the fiber.","core_discovery":"The central claim is that negative feedback between mark deposition and protein motion—a protein deposits an epigenetic mark and is repelled by it, which is impossible at thermodynamic equilibrium—converts ordinary diffusive sliding into unidirectional motion and allows proteins to penetrate collapsed chromatin. In the 1D approximation, marking a neighboring bead tilts the potential so the protein slides toward unmarked beads with rate $q_+ \\sim \\epsilon$ while backward hops are suppressed by a factor $e^{-\\epsilon/k_B T}$, yielding a run length $l_{\\rm run} \\sim e^{\\epsilon/k_B T}$. Multiple searchers repel each other through their epigenetic trails and spread hyperuniformly. In 3D simulations of a fiber collapsed by bridging proteins, an intermediate mark-evaporation rate $k_{\\rm off}$ produces a cusp in the fraction of the fiber visited per binding event, corresponding to searchers that locally open the globule and dive into its core. The authors conclude that chromophoresis is a generic nonequilibrium route to fast target search, potentially relevant to PARP-mediated lesion location.","pith_inferences":["Single-particle tracking of a marked-repelled protein on chromatin should reveal persistent, super-diffusive runs, a direct observational signature that the paper itself does not report.","Chemically slowing mark erasure should produce a non-monotonic change in lesion-search efficiency, a testable pharmacological prediction of the model.","Because the mechanism only requires a repulsive mark, proteins that deposit methylation or ubiquitylation marks could show the same running and diving dynamics, extending the PARP-centered biological framing.","The local chromatin swelling seen at DNA breaks may be caused by chromophoretic search itself rather than only by downstream signaling, which would make swelling a readout of search activity."],"forward_implications":["A repair protein meeting the $k_{\\rm on}$ threshold could locate a target on a human-sized chromosome in minutes rather than years, including targets buried inside collapsed heterochromatic globules.","Chromophoretic searchers are predicted to spread hyperuniformly, so each protein scans fresh chromatin rather than re-scanning marked trails, making search time scale favorably with searcher number.","An optimal mark-evaporation rate exists: too fast leaves the globule closed and too slow makes the fiber non-sticky, so cells could tune $k_{\\rm off}$ to control search speed, for example in response to DNA damage.","The mechanism is generic: any DNA-binding protein that deposits a repulsive mark should show the same running and diving behavior, broadening the relevance beyond repair to transcription and silencing."],"supporting_citations":[{"why":"Defines facilitated diffusion, the passive search strategy whose speed limit chromophoresis is claimed to exceed.","marker":"[13–16]"},{"why":"Supplies the effective 1D diffusion coefficient and the sliding model used to estimate the required mark-deposition rate.","marker":"[18]"},{"why":"Provides the microphase-separated polymer model of chromatin with epigenetic marks that the 3D simulations extend.","marker":"[21]"},{"why":"Introduces bridging-induced attraction, used to fold the polymer into a collapsed globule modelling heterochromatin.","marker":"[30]"},{"why":"Supplies the escape-rate theory for thermally activated hopping, from which the rates and run length are derived.","marker":"[35]"},{"why":"Gives measured rates for acetylation and phosphorylation, used to argue the required deposition rate is biologically plausible.","marker":"[37]"},{"why":"Documents that PARylation lowers chromatin-binding affinity and recruits repair enzymes, motivating negative chromophoresis.","marker":"[24, 25]"},{"why":"Shows PARylation swells chromatin in vitro, supporting the link between mark turnover and local unraveling.","marker":"[52]"}],"fun_headline_variants":["Fleeing own marks lets repair proteins dive into DNA","Nonequilibrium trick helps repair proteins find buried DNA faster","Proteins that repel own marks unzip collapsed DNA for fast repair","Self-repelling marks drive repair proteins into DNA cores","Fleeing their own tags, proteins slip into collapsed genome folds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The mechanism requires that a bound protein deposit its mark at a rate of at least about once per second ($k_{\\rm on} > D_{\\rm 1D}/\\sigma^2$); if real repair proteins modify histones more slowly than this, the unidirectional running state and the collapse-diving behavior would not occur in vivo.","fun_headline_variants_meta":{"raw":{"variants":["Fleeing own marks lets repair proteins dive into DNA","Nonequilibrium trick helps repair proteins find buried DNA faster","Proteins that repel own marks unzip collapsed DNA for fast repair","Self-repelling marks drive repair proteins into DNA cores","Fleeing their own tags, proteins slip into collapsed genome folds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000819,"raw_usage":{"total_tokens":3570,"prompt_tokens":914,"completion_tokens":2656,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":2585}},"tokens_in":530,"tokens_out":2656,"duration_ms":18140,"temperature":1.0,"reasoning_tokens":2585,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:37:11.398786+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track a single fluorescently labeled chromatin-binding protein as it deposits marks on a collapsed, bridging-protein-folded chromatin globule in vitro. The central claim predicts persistent unidirectional runs and repeated inward-diving events once the mark-deposition rate exceeds about 1 s$^{-1}$; observing only diffusive sliding and surface sticking at any deposition rate would disprove the chromophoretic mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the effective 1D diffusion coefficient and the sliding model used to estimate the required mark-deposition rate."},{"cited_title":"Michieletto, E","cited_arxiv_id":null,"evidence_quote":"Provides the microphase-separated polymer model of chromatin with epigenetic marks that the 3D simulations extend."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives measured rates for acetylation and phosphorylation, used to argue the required deposition rate is biologically plausible."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows PARylation swells chromatin in vitro, supporting the link between mark turnover and local unraveling."}],"review_version":1}