{"id":"e6370d18-48e1-4944-a001-1624bb7195a8","arxiv_id":"2502.06883","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"QuantumDNA packages LCAO-derived tight-binding parameters with Lindblad master equations to simulate DNA charge dynamics, reproducing known results and yielding preliminary exciton-lifetime maps for TERT promoter mutations.","lead":"QuantumDNA is an open-source Python package that simulates charge transfer and exciton dynamics in DNA, deriving tight-binding parameters from atomic structures and adding environmental effects through Lindblad equations. The paper demonstrates the package on published benchmarks and applies it to estimate how two melanoma-associated TERT promoter mutations might alter exciton lifetimes.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Accuracy claim is unvalidated for the core PDB-to-LCAO-to-TB workflow, since the benchmark examples bypass the default MSF parametrization and no independent comparison of LCAO parameters is provided.","rationale":"The reader's weakest_assumption identifies exactly the transferability of the MSF parameters. I agree, and add a sharper observation: the benchmarking section never validates the package's core workflow. Reproducing Bittner2007, Simserides2014, or Hawke2010 outputs demonstrates that the TB/Lindblad machinery is implemented, but it does not test the LCAO-derived parameters from the default parametrization. The TERT result is preliminary by the authors' own statement, so it cannot serve as validation. The package deserves credit for being open-source, documented, and modular, and the qualitative reproductions are useful, but the abstract's 'accurate' claim is conditional on an unverified transferability assumption. A comparison to reference transfer integrals would settle it. Therefore I support the reader's CONDITIONAL verdict; no further change is needed.","tokens_in":24077,"tokens_out":5266,"duration_ms":53497,"concrete_test":"Run QuantumDNA's calc_tb_params on the same B-DNA dimer geometries used in Mantela et al. 2021 and compare the resulting HOMO/LUMO energies and t_HOMO/t_LUMO to the published MSF/reference values; then repeat for the 14-bp TERT wild-type and C228T/C250T mutated sequences using the PDB-derived geometries. If the mean relative error exceeds about 10%, or if the sign of any nearest-neighbor transfer integral changes, the transferability assumption fails and the 'accurate' claim should be weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"QuantumDNA's headline promise is 'rapid yet accurate analysis' (Abstract). The accuracy of the core PDB-to-LCAO-to-TB workflow depends on the transferability of the Slater-Koster constants C_chi in Eqs. (2)-(3), fitted in the MSF parametrization [26] to a limited set of bases and ideal geometries, to arbitrary sequences, mutations, and PDB structures. This is not tested in the paper. Section 5.1 uses the Bittner2007 source; Section 5.2 uses Simserides2014 and is explicitly 'qualitative'; Section 5.3 runs the default Hawke2010 parameters. The only example that exercises the MSF LCAO path, the 1BNA GUI demo in Section 3.1, produces a population heatmap but no comparison to reference transfer integrals, orbital energies, or experimental rates. The TERT application is explicitly declared preliminary and not a validation. Nothing in the manuscript quantifies the error of the LCAO parameters against ab initio calculations or the experimental benchmarks the package claims to reproduce. Since every downstream observable (coherent dynamics, exciton lifetimes, mutation comparisons) inherits these parameters, an unvalidated transferability of C_chi is the load-bearing assumption. The concern is not that the package runs, but that the abstract's 'accurate' claim is unsupported: no evidence rules out, for example, errors of order 100 meV in t_HOMO or epsilon_HOMO for mutated bases.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents QuantumDNA, an open-source Python package that computes DNA charge-transfer and exciton-dynamics observables by combining an LCAO parameterization based on Slater-Koster two-center integrals (Section 4.1, Eqs. (2)-(3)) with tight-binding models at several resolutions (wire, ladder, extended ladder, fishbone variants, Fig. A2) and open-quantum-system dynamics via Lindblad master equations (Section 4.3, Eq. (10)). The package accepts PDB or XYZ geometries, offers a GUI, and is validated by three benchmark examples: the ultrafast excitonic dynamics after Bittner (Fig. 7), the superexchange-to-hopping crossover after Giese and Simserides (Fig. 8), and dephasing effects after Rossini et al. (Fig. 9), plus a preliminary biological application to the TERT promoter mutations C228T and C250T (Section 5.4, Fig. 10). The Abstract claims that the approach allows 'rapid yet accurate analysis of large DNA ensembles, enabling statistical studies of genetic and epigenetic phenomena.' The paper's central assertion is therefore that QuantumDNA is both efficient and quantitatively trustworthy for the PDB-to-LCAO-to-TB workflow it advertises.","tokens_in":24473,"tokens_out":10863,"duration_ms":102396,"significance":"The package fills a genuine niche: it appears to be the first documented, open-source toolkit that takes atomic-resolution DNA structures through LCAO-derived parameters and Lindblad dynamics to observables such as exciton lifetimes, dipole moments, and site-resolved populations, and it lowers the entry barrier with a GUI and tutorial notebooks. I explicitly credit the following strengths: the code is publicly available under a BSD-3 license with documentation at quantumdna.readthedocs.io and separate tutorial notebooks; the qualitative reproductions of Bittner's dynamics (Fig. 7) and of the Giese superexchange-hopping crossover (Fig. 8) are genuine evidence that the implementation runs and captures the expected physics; the high-throughput capability is demonstrated in the companion work Ref. [16] over all 16,384 seven-base-pair sequences; and the authors candidly state in Section 5.4 that the TERT results are 'not intended to validate' the biological effect.","major_comments":[{"comment":"The Abstract's 'rapid yet accurate' claim is unvalidated for the default MSF LCAO path, and the stress-test concern that the benchmarks bypass this path lands. None of the three benchmark examples in Section 5 exercises the default MSF parametrization: Section 5.1 uses the Bittner2007 source, Section 5.2 uses the Simserides2014 source and is explicitly 'qualitative' in both the Fig. 8 caption and the text, and Section 5.3 uses the Hawke2010/Rossini setup. The only example that runs the MSF path, the 1BNA GUI workflow in Section 3.1 and Fig. 4(d), outputs population heatmaps with no quantitative comparison to reference transfer integrals, orbital energies, or experimental observables. Because the constants C_chi are fitted to a limited set of bases and ideal geometries, their transferability to arbitrary sequences, mutated bases (as used in Section 5.4), and non-ideal PDB geometries is the load-bearing assumption behind the word 'accurate,' and the manuscript supplies no evidence that errors on the order of 100 meV in t_HOMO or epsilon_HOMO are excluded for mutated bases. I request a quantitative benchmark comparing MSF-derived t_HOMO, t_LUMO, epsilon_HOMO, and epsilon_LUMO with the ab initio and experimental values used in the fit of Ref. [26], plus a numerical error metric for the Fig. 7 reproduction against the original Bittner results [17].","section":"§4.1, Eqs. (2)-(3); §5.1-5.3; Abstract"},{"comment":"A large fraction of the validation is self-referential, which is a correctness risk rather than an accusation of misconduct: the default MSF parameterization (Ref. [26], Mantela et al.), the Simserides2014 benchmark source [53], and the Ref. [40] benchmark all originate from the present or closely overlapping authors, so agreement between package output and those published results can be inherited from shared code and fitted constants rather than from physical correctness. The only independently grounded element is the qualitative crossover of Fig. 8, eventually traceable to the Giese experiments [24]. A concrete test that would resolve the risk is to overlay the package's donor/acceptor population ratios with the measured rate ratios from Giese et al. [24] rather than only the crossover shape, or to compare MSF-LCAO parameters against independent electronic-structure calculations from groups not involved in the package, such as Kubař and Elstner [44] or Gutiérrez et al. [45, 46]. The paper currently does not provide the data needed for either check.","section":"§5, Figs. 8-9; Table A3"},{"comment":"The headline observables are controlled by ad hoc rates that are never anchored to data, and the TERT application in Section 5.4 inherits them. The exciton lifetime T is defined through the decay rate gamma_alpha in Eqs. (11)-(12), yet relax_rate = 3 rad/ps in the Section 3.1 example, deph_rate = 7 in Table A2, and loc_deph_rate = 2 in Section 5.3 are presented without justification against the measured excited-state lifetimes of Kohler et al. [18] or against the correlated dynamics of Ref. [16]. Since Fig. 10 plots differences in lifetime between mutant and natural sequences, those differences could be artifacts of the chosen gamma_alpha and of the free parameters J0, K0 in Eqs. (6)-(7); the paper itself concedes in Section 5.4 that the results are not intended to validate the biological effect. To make the TERT example evidential rather than illustrative, the authors should report absolute lifetimes, scan gamma_alpha and J0/K0 over a plausible range, and compare the resulting intervals with experimentally measured DNA excited-state lifetimes. In its present form, the TERT section supports 'the package can run' but not 'the package is accurate.'","section":"§4.4, Eqs. (11)-(12); §5.4, Fig. 10"}],"minor_comments":[{"comment":"The sentence 'as recently investigated by Kordas et al.' contains no citation; add the reference or remove the attribution.","section":"§3.1"},{"comment":"The justification for the nearest-neighbor Coulomb cutoff ('the thermal energy at 300 K... equals the Coulomb energy at the base-stacking distance D') implicitly uses the model's own J0, which is a free parameter whose default is 0 in Table A2; either derive the numerical comparison explicitly or present the cutoff as a modeling choice.","section":"§4.2.1, Eqs. (6)-(7)"},{"comment":"There is an apparent tension between the default 'parametrization' (MSF) and the default 'source' (Hawke2010), and between Table A3's statement that Mantela2021 is limited to the WM/base-pair level and the Section 3.1 demo, which generates ELM single-base parameters from the 1BNA PDB; clarify which parameterization actually produces the single-base ELM parameters in that demo.","section":"Table A2 and Table A3"},{"comment":"The index beta in Eq. (9) is never defined (presumably it labels the initial state), and the sentence 'the first term represents the time-averaged population for non-degenerate eigenenergies' glosses over the degenerate case; please define beta explicitly and state the degenerate-case formula or restrict the stated result.","section":"§4.2.2, Eq. (9)"},{"comment":"The plotted quantity is not fully specified: the caption says 'variation in lifetime' and 'differences between the lifetimes,' but the sign convention (mutant minus natural) and the color scale are not defined; please state them in the caption.","section":"§5.4, Fig. 10 caption"},{"comment":"The claim that Lindblad models are 'an efficient choice' is in tension with the scaling reported in Fig. A3 for the global thermalizing model (N^4 - N^2 operators); add a sentence noting the regime of applicability of each model.","section":"§4.3, Fig. A3"},{"comment":"There is a typo: 'amd' should be 'and'; also, the Abstract and the Introduction use different phrasings ('quantum physical methods' vs. 'quantum physics methods'), which should be harmonized.","section":"§1, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the paper is a software/package paper, and the journal should weigh whether its validation norms require at least one quantitative benchmark independent of the author group; Majors 1 and 2 together ask for exactly that. The citation pattern is a point to be aware of but not, by itself, a problem: the default MSF parameterization (Ref. [26]) and benchmark sources Refs. [40] and [53] overlap with the author list, so an editor may wish to check that the version of record includes the code version and reproducibility statement referenced in the Data Availability section."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely useful open-source package paper, not a new-physics paper, and the main thing to know is that the \"accurate\" in the abstract is currently more promise than demonstrated property. The code is public, modular, and reproduces known qualitative results, which is real value for a field that keeps reimplementing the same tight-binding models by hand. Credit where due: the package bundles LCAO, TB, Lindblad, a GUI, and benchmark notebooks; the TERT mutation maps are new outputs, and the authors are careful to call them preliminary. That care is good practice.\n\nThe soft spot is the one the stress-test names. Every benchmark that actually validates the code bypasses the default MSF LCAO path: the Bittner reproduction uses the Mehrez-Anantram parameters, the Simserides reproduction is explicitly qualitative, and the dephasing example uses Hawke2010. The only PDB-to-LCAO demo (1BNA) shows a heatmap but no comparison to reference transfer integrals or orbital energies. So the abstract's \"rapid yet accurate\" is not supported for the workflow the package advertises most prominently. If the Slater-Koster constants C_chi do not transfer to mutated bases or PDB geometries, the TERT results and every similar output inherit that error. This is not a fatal flaw—the software works—but the claim needs either quantitative benchmarks or a softer abstract.\n\nAlso worth noting: two of the three benchmarks are the authors' own prior work. That is fine when the point is reimplementation, but it makes the independent confirmation thinner than it looks. Minor point: the ad hoc rates (relax_rate=3, deph_rate=7) get no sensitivity analysis, and the lifetime definition with epsilon=1/e is fine but should be stated more prominently.\n\nWho this is for: someone doing DNA charge-transfer simulations who wants a working, documented baseline instead of writing their own tight-binding code. That reader gets value now, but should treat absolute numbers with caution.\n\nVerdict: this deserves serious peer review. A software paper with public code, reproductions, and an honest limitations section is exactly what referees should spend time on. The referee should ask for quantitative comparisons or a trimmed abstract, but this is not a desk reject.","headline":"Useful, honest software paper whose 'accurate' claim outruns its validation.","tokens_in":24966,"tokens_out":1786,"would_cite":true,"duration_ms":19053,"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":"QuantumDNA maps DNA structures to charge-transfer predictions","keywords":["DNA charge transfer","tight-binding models","LCAO Slater-Koster method","Lindblad master equation","exciton dynamics","open quantum systems","Python package","TERT promoter mutations"],"falsifier":"Compute the interbase transfer integrals $t_{AB}$ with a higher-level electronic-structure method for the same atomic coordinates used by QuantumDNA in a mutated or distorted case, such as the TERT C228T or C250T sequence, and compare with the package's LCAO values; systematic deviations large enough to change the predicted exciton lifetime beyond the model's spread would falsify the transferability claim. A pump-probe experiment on a mutated versus wild-type oligonucleotide could then test the predicted lifetime shift directly.","tokens_in":23869,"feed_emoji":"🧬","tokens_out":8916,"duration_ms":79296,"temperature":0.7,"pith_summary":"QuantumDNA is an open-source Python package that simulates charge transfer and excited-state dynamics along DNA. The paper's central claim is that a pipeline combining LCAO-derived electronic couplings, coarse-grained tight-binding models, and Lindblad master equations can turn atomic DNA structures into quantitative predictions of hole, electron, and exciton motion quickly enough for large-scale sequence screening. The authors validate the pipeline by reproducing published results on ultrafast exciton dynamics, the crossover from tunneling to thermally activated hopping, and environment-induced decoherence. They then apply it to two melanoma-associated TERT promoter mutations and find that C228T shortens exciton lifetimes on one strand and lengthens them on the other, while C250T generally lengthens lifetimes, while explicitly labeling these findings preliminary. If the accuracy claim holds, the package gives biologists and clinicians a reusable bridge from DNA sequence and structure to quantum-dynamics observables.","feed_headline":"QuantumDNA predicts how DNA mutations shift exciton lifetimes","feed_subtitle":"Open-source package combines LCAO couplings, tight-binding models, and Lindblad noise for fast DNA screening.","key_machinery":"The load-bearing object is the LCAO computation of interbase couplings: molecular orbitals are built as linear combinations of valence atomic orbitals, and the couplings are Slater-Koster two-center transfer integrals with a quadratic decay inside a base and an exponential decay, with $d_0 = 1.35$ Å, between bases, using fitted constants $C_\\chi$ from the default MSF parametrization. These couplings parameterize tight-binding Hamiltonians at several resolutions, from wire to extended-ladder to fishbone models. Environmental effects enter through Lindblad operators for local or global dephasing, thermalization, and site-wise exciton recombination into the ground state. The key observables are the exciton lifetime, defined as the first time at which the ground-state population reaches $1 - 1/e$, and the mean electron-hole separation, obtained from a dipole operator. This chain from geometry to Hamiltonian to master equation to observable is what carries the argument.","core_discovery":"The central claim is that a single software platform can integrate the whole chain: Slater-Koster two-center integrals evaluated from atomic coordinates, a family of coarse-grained tight-binding models, Lindblad dissipators for dephasing, thermalization, and exciton recombination, and observables such as exciton lifetime and electron-hole separation. The paper asserts that this integration is accurate enough to reproduce published benchmarks and fast enough to scan all 16,384 seven-base-pair sequences, as was done in prior work. The TERT promoter example is presented as a first biological application: mutation-dependent exciton-lifetime changes appear near the mutation site, but the authors caution that the results do not validate a quantum influence on melanoma mechanisms.","pith_inferences":["If the fitted Slater-Koster constants transfer to arbitrary sequences and mutated geometries, the package could serve as a pre-filter that ranks DNA regions by electronic response before costlier calculations; the paper does not itself establish that transferability.","The single-exciton constraint keeps the Hilbert space quadratic in chain length, so the results apply to dilute photoexcitation; intense excitation or multi-exciton processes are out of scope and would need a larger state space.","The TERT lifetime pattern suggests a testable experimental prediction: short mutated oligonucleotides with C228T or C250T should show strand-dependent lifetime shifts in pump-probe experiments, although the paper presents the pattern as exploratory.","Quantitative biological conclusions will also depend on environmental parameters like dephasing and relaxation rates; the defaults are a convenient starting point rather than a validated model of the cellular environment."],"forward_implications":["A user can go from a PDB or XYZ structure to simulated transfer integrals, populations, exciton lifetimes, and dipole moments with a few lines of Python or through the GUI.","Published one-off models, including ultrafast exciton dynamics, the tunneling-to-hopping crossover, and dephasing-induced coherence loss, become reproducible building blocks inside one package.","Sequence ensembles large enough for statistical studies, such as all 16,384 seven-base-pair sequences, can be screened for electronic properties rather than studied one at a time.","Targeted mutations in real structures, such as the TERT promoter mutations, can be screened for their effect on exciton lifetimes before expensive ab initio or experimental follow-up.","The choice among wire, ladder, extended-ladder, and fishbone models lets users trade resolution against computational cost for the system at hand."],"supporting_citations":[{"why":"Supplies the default MSF LCAO parametrization with all valence orbitals that fixes the fitted constants $C_\\chi$.","marker":"[26]"},{"why":"Introduces the LCAO and Slater-Koster route for DNA transfer integrals that QuantumDNA's parameter calculator builds on.","marker":"[33]"},{"why":"Provides the refined single-base and base-pair parameter set used as a built-in source and the exponential interbase decay form.","marker":"[29]"},{"why":"Supplies the exciton tight-binding model with Coulomb and exchange interactions that Section 5.1 reproduces.","marker":"[17]"},{"why":"Extends the Frenkel-exciton modeling of AT DNA that the package's exciton benchmarks draw on.","marker":"[32]"},{"why":"Supplies the ab initio transfer parameters behind the Bittner benchmark source.","marker":"[51]"},{"why":"Established the high-throughput exciton-lifetime pipeline for all 16,384 seven-base-pair sequences that QuantumDNA makes reproducible.","marker":"[16]"},{"why":"Defines the local versus global dephasing scenarios and coherence analysis reproduced in Section 5.3.","marker":"[40]"},{"why":"Provides the wire-model parameter set and the superexchange-to-hopping crossover reproduced in Section 5.2.","marker":"[53]"},{"why":"Provides the experimental basis for long-lived excited states and the relaxation channel modeled by the package.","marker":"[18]"}],"fun_headline_variants":["QuantumDNA: open-source simulations of DNA charge dynamics and mutations","Predicting mutation effects on exciton lifetimes with QuantumDNA","QuantumDNA unifies methods to predict DNA mutation effects on charge","Scan DNA mutation impacts on exciton lifetimes fast with QuantumDNA","QuantumDNA: fast, open-source DNA charge dynamics with GUI"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire accuracy claim rests on the fitted Slater-Koster constants $C_\\chi$ remaining valid for arbitrary sequences, mutated bases, and PDB-derived geometries, and the paper gives no independent check of that transferability.","fun_headline_variants_meta":{"raw":{"variants":["QuantumDNA: open-source simulations of DNA charge dynamics and mutations","Predicting mutation effects on exciton lifetimes with QuantumDNA","QuantumDNA unifies methods to predict DNA mutation effects on charge","Scan DNA mutation impacts on exciton lifetimes fast with QuantumDNA","QuantumDNA: fast, open-source DNA charge dynamics with GUI"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001333,"raw_usage":{"total_tokens":5370,"prompt_tokens":843,"completion_tokens":4527,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":459,"completion_tokens_details":{"reasoning_tokens":4442}},"tokens_in":459,"tokens_out":4527,"duration_ms":33230,"temperature":1.0,"reasoning_tokens":4442,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T18:32:15.799659+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the interbase transfer integrals $t_{AB}$ with a higher-level electronic-structure method for the same atomic coordinates used by QuantumDNA in a mutated or distorted case, such as the TERT C228T or C250T sequence, and compare with the package's LCAO values; systematic deviations large enough to change the predicted exciton lifetime beyond the model's spread would falsify the transferability claim. A pump-probe experiment on a mutated versus wild-type oligonucleotide could then test the predicted lifetime shift directly.","supporting_citations":[{"cited_title":"Mantela, C","cited_arxiv_id":null,"evidence_quote":"Supplies the default MSF LCAO parametrization with all valence orbitals that fixes the fitted constants $C_\\chi$."},{"cited_title":"Electronic properties of DNA: structural and chemical influence on the quest for high conductance and charge transfer","cited_arxiv_id":"cond-mat/0201404","evidence_quote":"Introduces the LCAO and Slater-Koster route for DNA transfer integrals that QuantumDNA's parameter calculator builds on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the exciton tight-binding model with Coulomb and exchange interactions that Section 5.1 reproduces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Extends the Frenkel-exciton modeling of AT DNA that the package's exciton benchmarks draw on."},{"cited_title":"Mehrez, M","cited_arxiv_id":null,"evidence_quote":"Supplies the ab initio transfer parameters behind the Bittner benchmark source."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Established the high-throughput exciton-lifetime pipeline for all 16,384 seven-base-pair sequences that QuantumDNA makes reproducible."},{"cited_title":"Effect of environmental noise on charge diffusion in DNA: Towards modeling its potential epigenetic impact in live processes","cited_arxiv_id":"2407.14252","evidence_quote":"Defines the local versus global dephasing scenarios and coherence analysis reproduced in Section 5.3."},{"cited_title":"Simserides, A systematic study of electron or hole transfer along DNA dimers, trimers and polymers, Chemical Physics 440 (2014) 31–41","cited_arxiv_id":null,"evidence_quote":"Provides the wire-model parameter set and the superexchange-to-hopping crossover reproduced in Section 5.2."},{"cited_title":"Crespo-Hernandez, B","cited_arxiv_id":null,"evidence_quote":"Provides the experimental basis for long-lived excited states and the relaxation channel modeled by the package."}],"review_version":1}