Pith. sign in

REVIEW 4 major objections 4 minor 48 references

Thermally Robust Cooperative Function in Mechanical Networks

T0 review · 4 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Elastic networks trained for symmetric bidirectional allostery can exhibit cooperative binding that remains functional at biologically relevant temperatures.

desk verdict Good thermal-fluctuation framework for trained networks, but the 300 K robustness claim depends on a lenient 1-sigma criterion and a strain calibration that is a judgment call. read the letter →

arxiv 2509.20560 v2 pith:QALQ3OJM submitted 2025-09-24 cond-mat.soft

Ben Pisanty , Jovana A. Kim , Andrea J. Liu , Sidney R. Nagel This is my paper · ORCID
classification cond-mat.soft
keywords cooperativebindingallosteryelasticnetworksthermalfluctuationscrossovertemperaturecoupledlearningmechanicaltrainingproteinmodels
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper shows that a disordered spring network can be trained, by applying allosteric training symmetrically at two sites, so that straining either site lowers the energy cost of straining the other—the signature of cooperative binding. The authors derive a crossover temperature T* at which thermal fluctuations at one site are as large as the signal from the other, and they show that T* is non-monotonic in the trained response amplitude and vanishes at the fully cooperative limit. They then modify the training cost function to selectively stiffen the low-frequency mode that carries the cooperative response, which raises T*. Using a Lindemann criterion for melting, they find a window of training parameters in which both T* and the melting temperature lie above 300 K, their proxy for biologically relevant conditions. The authors present the result as a proof of principle that thermal stability of cooperative function is achievable in model elastic networks.

What carries the argument

The load-bearing object is the cooperative mode, the low-frequency normal mode onto which the symmetric strain response projects. Because the thermal strain variance at a site is a sum over modes of kBT/(mω²) times the mode's projection onto that site, a soft mode both enables large response and invites large thermal noise. The training modification is a generalized contrast function C̃ = (1−r)(EC−EF) − rEF, which for small r preferentially stiffens the bonds that store energy in the free state, raising the functional mode's frequency while preserving the trained response. This trade-off is what opens the window where both T* and TM exceed 300 K.

What would settle it

Using elastic-network normal modes of real cooperative proteins (e.g., hemoglobin), compute T* from Eq. 6 with measured source strains; if the resulting T* values cluster below 300 K for proteins that are known to cooperate at that temperature, the energetic mechanism proposed here cannot account for protein robustness.

Watch

Extended reading notes

Core claim

Cooperative binding in a mechanical network can be trained as bidirectional allostery: the same allosteric response amplitude Δ is imposed from site A to B and from B to A. The training uses a symmetric version of the coupled-learning contrast rule, and the resulting function is carried by a single low-frequency vibrational mode. The paper derives the one-sigma thermal strain fluctuation at a target site from normal modes (Eq. 6) and defines the crossover temperature T* as the point where that fluctuation equals the trained mean response. A generalized contrast function with parameter r trades off a small amount of training accuracy for a higher free-state energy, which raises the frequency

Load-bearing premise

Because training is done at zero temperature, the cooperative mechanism is purely energetic; if real protein cooperativity involves significant entropic effects, prestress, or frustration, the thermal stability measured here may not transfer to proteins.

Editorial extensions

If this is right

  • Cooperative binding can be trained as a symmetric allosteric task, and it produces higher cooperativity than unidirectional allostery training, with the cooperativity parameter R approaching 1 as Δ→1.
  • The crossover temperature is non-monotonic in the response amplitude: it rises with Δ for small Δ, then falls as the cooperative mode softens, vanishing at both Δ=0 and Δ=1.
  • A small penalty for free-state energy (r up to about 5×10⁻⁶) raises the dimensionless crossover temperature; larger r stiffens less important bonds and lowers Θ* while still raising the dimensional T*.
  • There exist trained networks for which both the functional crossover temperature and the melting temperature lie above the 300 K protein-equivalent threshold, so thermal robustness is achievable in principle.
  • The linear-response prediction of strain fluctuations, Eq. 6, matches molecular dynamics simulation quantitatively even beyond the linear regime and up to two orders of magnitude above T*.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If evolution can tune protein stiffness distributions the way this training does, the same trade-off between functional-mode frequency and overall softness may explain why some cooperative proteins are more thermally stable than others.
  • The crossover criterion based on 1-σ fluctuations is permissive; the SI's alternative that includes bound-state fluctuations gives lower T*, so real proteins may need even larger margins than the reported window.
  • Uniform stiffening of all springs also raises the dimensional T*, but the paper's dimensionless analysis implies that the non-trivial design rule is to concentrate stiffness in the functional mode—a testable principle for synthetic metamaterials.
  • The entropic-mechanism caveat suggests a concrete follow-up: train networks at finite temperature or with prestress, and compare the resulting T* to the T=0-trained case; if finite-temperature training yields lower T*, the energetic mechanism is not sufficient for naturally evolved proteins.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper trains disordered elastic spring networks at zero temperature for cooperative (bidirectional) allostery, defines a crossover temperature T* above which thermal fluctuations destroy the trained response, and introduces a modified contrast function with parameter r that raises T* by stiffening the relevant low-frequency mode. The authors validate their linear-response strain-fluctuation calculation against LAMMPS simulations, compute a Lindemann-type melting temperature, and compare both T* and T_M to a 300 K threshold using protein-derived estimates of strain, length, and stiffness. They conclude that for a range of (Δ, r) both temperatures lie above biologically relevant values, so cooperative function can be thermally robust.

Significance. The linear-response framework (Eqs. 6, 17) is standard and the validation against MD in the SI (average ~6% error up to T*) is a real strength. The idea that training for bidirectional allostery yields cooperativity, and that the associated soft mode controls thermal robustness, is plausible and potentially useful for designing synthetic mechanical networks. The modified training protocol with r is a clean extension. However, the headline claim of robustness at 300 K depends on a chain of dimensional conversions and criterion choices that are not fully justified; the internal mechanics are sound, but the biological extrapolation is sensitive to the chosen binding strain and to the 1-sigma crossover definition. The authors explicitly acknowledge that their mechanisms are energetic and not entropic, which is a genuine limitation for protein relevance.

major comments (4)
  1. [SI, 'Median Value for Source Strain εs'; Fig. S19] The central 300 K comparison in Fig. 4 depends on the representative source strain εs through Eq. 8, where T* scales as ε_s^2. The SI selects εs=0.19 by maximizing U(p)=M(p)/(W(p)+1/M'(p)), yielding p*≈83% of the intra-protein atom-pair strain distribution. This is not a physical argument that a coarse-grained bond strain maps to the 83rd percentile of atom-pair strains. The authors' own 25th–75th percentile range is 0.09–0.31; using 0.09 increases the 300 K threshold by (0.19/0.09)^2≈4.5, which can remove the circled points in Fig. 4. Please provide a robustness analysis or a more principled determination.
  2. [Eq. (8), Sec. III; SI, 'Alternative Definition for Crossover Condition'] Eq. (8) defines T* as the temperature at which the 1-sigma unbound strain fluctuation equals the mean bound strain. Calling this 'the crossover temperature above which functionality breaks down' is a criterion choice, not a physical phase transition. Requiring 2-sigma or 3-sigma separation reduces T* by factors of 4 or 9; the alternative in SI Eq. 4 is still a 1-sigma overlap criterion and is lower. The paper should report T* for at least two separation thresholds and show how the Fig. 4 window depends on this choice, because the claim of biological robustness is sensitive to it.
  3. [Sec. VII C and SI Eq. (3)] The dimensionless-to-real temperature conversion uses k∼(L/5)Y with Y≈10 GPa and L≈5 nm as fixed values. The gray bands in Figs. 2–4 propagate only the inter-protein variability in L and ε_s (or L_M), not the uncertainty in Y or in the effective spring stiffness. Since Θ in SI Eq. 3 is inversely proportional to Y and L^3, a factor of 2 uncertainty in Y shifts the 300 K line substantially. Please propagate these uncertainties or justify the fixed values with cited ranges; this affects whether the claimed (Δ,r) window survives.
  4. [Sec. VI, Discussion] The Discussion states that because training is at T=0, all finite-temperature effects are energetic fluctuations about the trained state, and that entropic mechanisms, prestress, or frustration are not included. This is an explicitly acknowledged limitation, and it is load-bearing for the extrapolation to real protein cooperativity (though not for the internal network result). The conclusions should state more cautiously that the biological relevance claim applies to energetic mechanisms only, and should identify a concrete test—e.g., training with prestress or a finite-temperature cost function—that could distinguish the two regimes.
minor comments (4)
  1. [SI, various] Typos and wording issues: 'Dimentional' (SI headings), 'T o' in Fig. S17 caption, 'strick' in the εs section, 'yeilding' in the same section. Please proofread.
  2. [Fig. 2D / Fig. 3 captions] The gray dashed line and shaded band are described as '300 K' but they are estimates based on several model assumptions. Consider labeling them as '300 K estimate (median/quartiles)' and noting the conversion in every caption where they appear.
  3. [SI Eq. (3) and Eq. (8)] The symbol q in SI Eq. (3) is defined as ε for Θ* and L_M for ΘM, while the main text uses ε_s in Eq. (8). Use one consistent notation (e.g., ε_s always) to avoid confusion.
  4. [Sec. II, Eq. (4)] The definition of Ecoop is clear, but the sign convention for EA|B vs EB|A is not obvious from the text alone. A sentence explicitly stating that lower EA|B corresponds to facilitation would help.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the crossover and melting temperatures are computed from the trained network's own spectrum and independently calibrated protein parameters, with MD validation.

full rationale

The derivation chain is self-contained. The crossover temperature is introduced via Eq. 8 as the temperature at which the one-sigma thermal strain fluctuation in the unbound state equals the trained mean response; although this is a definitional criterion, the fluctuation magnitude is not fitted to any target T* or to 300 K—it is computed from the normal-mode spectrum (Eq. 6) and validated against explicit MD simulations (SI Figs. S1-S2). The melting temperature follows from the standard Lindemann criterion with literature values L_M in 0.1-0.2, again not fitted to the paper's conclusion. The conversion to real temperatures uses independently estimated protein parameters (Y ~ 10 GPa, L ~ 5 nm, PDB-derived binding strains); the SI's percentile-based selection of epsilon_s is a parameter-estimation choice with acknowledged inter-protein spread, not a quantity fitted to make Fig. 4 cross the 300 K line. Self-citations (Refs. 15, 20, 24, 25) are used for methodology or for the known low-frequency-mode phenomenology, and are supported either by the paper's own Fig. 1E-F or by an independent energy estimate (epsilon_s ~ 0.2). The stated limitation that training is T=0 and energetic in origin is a transferability caveat, not a circular step. No equation is equivalent to its input by construction, and no fitted parameter is relabeled as a prediction.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The ledger captures the empirical inputs for the protein temperature mapping (strains, stiffness, lengths, Lindemann threshold) and the modeling assumptions that connect trained spring networks to cooperative binding. The crossover criterion is a definition, not a derived transition, and is listed as an ad hoc axiom. No new physical entities are introduced.

free parameters (6)
  • r (contrast-function trade-off) = scanned: 0, 5e-7, 1e-6, 2e-6, 5e-6, 1e-5, 5e-5; optimal r about 5e-6
    Introduced in Eq. 9 to control trade-off between cooperative function and raising free-state energy. The paper scans r rather than fitting it; the optimal value emerges from the stiffness bounds.
  • eta (nudge factor) = 0.01
    Standard coupled learning hyperparameter; chosen small; affects the learning rule Eq. 3 and the constraint r << eta*error.
  • source strain epsilon_s = median 0.19, range 0.09-0.31
    Estimated from PDB analysis of 24 allosteric proteins (SI, percentile method). Used to convert dimensionless crossover temperature to real temperature; not fitted to the target result.
  • effective spring stiffness k = ~10 N/m
    Derived from protein Young's modulus Y~10 GPa and bond length ~1 nm (l = L/5 with L~5 nm). Used in temperature conversion.
  • protein length L = median 5.4 nm, range 4.7-6.6 nm
    Measured as maximal distance between atoms in 24 allosteric proteins; used in mapping.
  • Lindemann parameter LM = 0.15 (range 0.1-0.2)
    Taken from melting literature for proteins; used to define melting temperature and the 300K band.
assumptions (6)
  • domain assumption The network is a valid coarse-grained model of protein allostery/cooperativity; binding is modeled as prescribed strain at a pair of nodes.
    Throughout the paper, trained spring networks are used as models for protein allostery, following Ref. [13] etc. The mapping to proteins relies on this.
  • domain assumption The elastic energy required to strain a binding site is a valid proxy for the binding energy barrier (Eq. 4).
    Used to define Ecoop and cooperativity R; if the proxy fails, the functional interpretation of the trained response as cooperative binding is unsupported.
  • standard math Equipartition and normal-mode linear response describe finite-temperature strain fluctuations (Eqs. 6, 15-17).
    The crossover temperature derivation relies on Gaussian displacement coefficients with variance kBT/(m omega^2). Validated against MD in SI.
  • ad hoc to paper The crossover criterion: function is destroyed when 1-sigma unbound target strain fluctuations equal the bound mean strain (Eq. 8).
    T* is defined by this criterion; an alternative definition in SI gives qualitatively similar but quantitatively different results. The claim of biological robustness is sensitive to this definition.
  • domain assumption Mechanisms underlying cooperative function are fundamentally energetic; entropic contributions are neglected.
    Stated explicitly in Discussion: training at T=0 yields energetic mechanisms. The finite-temperature behavior is described by thermal fluctuations around the T=0 trained state.
  • domain assumption Lindemann criterion applies to proteins with LM in 0.1-0.2.
    Used to compute melting temperature and the 300K gray band.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Thermally Robust Cooperative Function in Mechanical Networks." pith.science (2026). https://pith.science/paper/QALQ3OJM

@misc{pith2026250920560,
  author       = {Pith},
  title        = {Pith review of: Thermally Robust Cooperative Function in Mechanical Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QALQ3OJM}},
  note         = {Machine review of arXiv:2509.20560}
}
read the original abstract

Elastic networks can be tuned to exhibit complex mechanical responses and have been extensively used to study protein allostery, where a localized strain regulates the conformation at a distant site. We show that cooperative binding, where one site enhances the other's ability to function, can be improved by tuning for \textit{bidirectional} allostery: a symmetric coupling where the strain propagated between the two sites is independent of which site is perturbed. We identify a crossover temperature above which functionality breaks down due to thermal fluctuations. We introduce a modified tuning process to increase this crossover temperature, showing that function can be robust at biologically relevant temperatures if thermal stability is important.

Figures

Figures reproduced from arXiv: 2509.20560 by the authors.

Figure 1
Figure 1. FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 5 canonical work pages

  1. [1]

    Monod, J

    J. Monod, J. Wyman, and J.-P. Changeux, Journal of Molecular Biology 12, 88 (1965)

    1965

  2. [2]

    Changeux, Protein Science 20, 1119 (2011)

    J.-P. Changeux, Protein Science 20, 1119 (2011)

    2011

  3. [3]

    un- bound

    Filled circles indicate that, in addition to both being abo ve the 300 K threshold, the crossover temperature is also larger than the melting temperature. Fig. S15). For a given ∆, however, the system stiffens with increasing r, resulting in an increasing TM with r (Fig. S15). The dimensionless melting temperature Θ M , however, consistently decreases with...

  4. [4]

    Tsai and R

    C.-J. Tsai and R. Nussinov, PLOS Computational Biol- ogy 10, 1 (2014)

    2014

  5. [5]

    S. J. Wodak, E. Paci, N. V. Dokholyan, I. N. Berezovsky, A. Horovitz, J. Li, V. J. Hilser, I. Bahar, J. Karani- colas, G. Stock, P. Hamm, R. H. Stote, J. Eberhardt, Y. Chebaro, A. Dejaegere, M. Cecchini, J.-P. Changeux, P. G. Bolhuis, J. Vreede, P. Faccioli, S. Orioli, R. Rava- sio, L. Yan, C. Brito, M. Wyart, P. Gkeka, I. Rivalta, G. Palermo, J. A. McCam...

    2019

  6. [6]

    M. M. Tirion, Phys. Rev. Lett. 77, 1905 (1996)

    1905

  7. [7]

    Bahar, A

    I. Bahar, A. R. Atilgan, and B. Erman, Folding and De- sign 2, 173 (1997). 9

    1997

  8. [8]

    Haliloglu, I

    T. Haliloglu, I. Bahar, and B. Erman, Phys. Rev. Lett. 79, 3090 (1997)

    1997

Show all 48 references
  1. [9]

    Hinsen, Proteins: Structure, Function, and Bioinfor- matics 33, 417 (1998)

    K. Hinsen, Proteins: Structure, Function, and Bioinfor- matics 33, 417 (1998)

    1998

  2. [10]

    A. J. Pak and G. A. Voth, Current Opinion in Structural Biology 52, 119 (2018), cryo electron microscopy: the impact of the cryo-EM revolution in biology • Biophysical and computational methods - Part A

    2018

  3. [11]

    Flechsig and A

    H. Flechsig and A. S. Mikhailov, Journal of The Royal Society Interface 16, 20190244 (2019), https://royalsocietypublishing.org/doi/pdf/10.1098/rsif.2019.0244

    2019

  4. [12]

    M. R. Mitchell, T. Tlusty, and S. Leibler, Proceedings of the National Academy of Sciences 113, E5847 (2016), https://www.pnas.org/doi/pdf/10.1073/pnas.1609462113

    2016 doi

  5. [13]

    Flechsig, Biophysical Journal 113, 558 (2017)

    H. Flechsig, Biophysical Journal 113, 558 (2017)

    2017

  6. [14]

    J. W. Rocks, N. Pashine, I. Bischofberger, C. P. Goodrich, A. J. Liu, and S. R. Nagel, Proceedings of the National Academy of Sciences 114, 2520 (2017), https://www.pnas.org/doi/pdf/10.1073/pnas.1612139114

    2017 doi

  7. [15]

    L. Yan, R. Ravasio, C. Brito, and M. Wyart, Proceedings of the National Academy of Sciences 114, 2526 (2017), https://www.pnas.org/doi/pdf/10.1073/pnas.1615536114

    2017 doi

  8. [16]

    L. Yan, R. Ravasio, C. Brito, and M. Wyart, Biophys J 114, 2787 (2018)

    2018

  9. [17]

    Ravasio, S

    R. Ravasio, S. M. Flatt, L. Yan, S. Zamuner, C. Brito, and M. Wyart, Biophys J 117, 1954 (2019)

    1954

  10. [18]

    J. W. Rocks, H. Ronellenfitsch, A. J. Liu, S. R. Nagel, and E. Katifori, Proceedings of the Na- tional Academy of Sciences 116, 2506 (2019), https://www.pnas.org/doi/pdf/10.1073/pnas.1806790116

    2019 doi

  11. [19]

    Bravi, R

    B. Bravi, R. Ravasio, C. Brito, and M. Wyart, PLOS Computational Biology 16, 1 (2020)

    2020

  12. [20]

    Rouviere, R

    E. Rouviere, R. Ranganathan, and O. Rivoire, PRX Life 1, 023004 (2023)

    2023

  13. [21]

    J. W. Rocks, E. Katifori, and A. J. Liu, Topological characterization of the continuum of allosteric response (2024), arXiv:2401.13861 [cond-mat.soft]

    2024 arXiv

  14. [22]

    M. M. Rubin and J.-P. Changeux, Journal of Molecular Biology 21, 265 (1966)

    1966

  15. [23]

    M. I. Stefan and N. Le Nov` ere, PLOS Computational Biology 9, 1 (2013)

    2013

  16. [24]

    Loutchko and H

    D. Loutchko and H. Flechsig, Biophysical Reviews 12, 443 (2020)

    2020

  17. [25]

    Stern, A

    M. Stern, A. J. Liu, and V. Balasubramanian, Phys. Rev. E 109, 024311 (2024)

    2024

  18. [26]

    Stern, M

    M. Stern, M. Guzman, F. Martins, A. J. Liu, and V. Balasubramanian, Physical networks become what they learn (2024), arXiv:2406.09689 [cond-mat.dis-nn]

    2024 arXiv

  19. [27]

    C. P. Goodrich, A. J. Liu, and S. R. Nagel, Phys. Rev. Lett. 114, 225501 (2015)

    2015

  20. [28]

    Pashine, D

    N. Pashine, D. Hexner, A. J. Liu, and S. R. Nagel, Science Advances 5, eaax4215 (2019), https://www.science.org/doi/pdf/10.1126/sciadv.aax4215

    2019 doi

  21. [29]

    Stern, D

    M. Stern, D. Hexner, J. W. Rocks, and A. J. Liu, Phys. Rev. X 11, 021045 (2021)

    2021

  22. [30]

    Stern, S

    M. Stern, S. Dillavou, M. Z. Miskin, D. J. Durian, and A. J. Liu, Phys. Rev. Res. 4, L022037 (2022)

    2022

  23. [31]

    Moritsugu and J

    K. Moritsugu and J. C. Smith, Biophysical Journal 93, 3460 (2007)

    2007

  24. [32]

    Lyman, J

    E. Lyman, J. Pfaendtner, and G. A. Voth, Biophysical Journal 95, 4183 (2008)

    2008

  25. [33]

    L. Yang, G. Song, and R. L. Jernigan, Proceedings of the National Academy of Sciences 106, 12347 (2009), https://www.pnas.org/doi/pdf/10.1073/pnas.0902159106

    2009 doi

  26. [34]

    Ming and M

    D. Ming and M. E. Wall, Phys. Rev. Lett. 95, 198103 (2005)

    2005

  27. [35]

    Dubanevics and T

    I. Dubanevics and T. C. McLeish, Journal of Molecular Biology 434, 167696 (2022), allostery: From Mechanisms to Therapies

    2022

  28. [36]

    J. R. Gissinger, I. Nikiforov, Y. Afshar, B. Waters, M.- k. Choi, D. S. Karls, A. Stukowski, W. Im, H. Heinz, A. Kohlmeyer, and E. B. Tadmor, The Journal of Phys- ical Chemistry B 128, 3282 (2024), pMID: 38506668, https://doi.org/10.1021/acs.jpcb.3c08419

    2024 doi

  29. [37]

    Sheinman, C

    M. Sheinman, C. P. Broedersz, and F. C. MacKintosh, Phys. Rev. E 85, 021801 (2012)

    2012

  30. [38]

    F. A. Lindemann, The calculation of molecular Eigen- frequencies, Tech. Rep. (1984)

    1984

  31. [39]

    F. H. Stillinger and D. K. Stillinger, The Journal of Chemical Physics 93, 6013 (1990), https://pubs.aip.org/aip/jcp/article- pdf/93/8/6013/18989431/6013 1 online.pdf

    1990

  32. [40]

    Y. Zhou, D. Vitkup, and M. Karplus, Journal of Molec- ular Biology 285, 1371 (1999)

    1999

  33. [41]

    J. L. S. Eshel Faraggi, A. Keith Dunker and A. Kloczkowski, Journal of Biomolecular Structure and Dynamics 36, 2331 (2018), pMID: 28714803, https://doi.org/10.1080/07391102.2017.1352539

    2018 arXiv

  34. [42]

    Katava, G

    M. Katava, G. Stirnemann, M. Zanatta, S. Ca- paccioli, M. Pachetti, K. L. Ngai, F. Ster- pone, and A. Paciaroni, Proceedings of the Na- tional Academy of Sciences 114, 9361 (2017), https://www.pnas.org/doi/pdf/10.1073/pnas.1707357114

    2017 doi

  35. [43]

    T. Ku, P. Lu, C. Chan, T. Wang, S. Lai, P. Lyu, and N. Hsiao, Computational Biology and Chemistry 33, 445 (2009)

    2009

  36. [44]

    H. P. Erickson, Biological Procedures Online 11, 32 (2009)

    2009

  37. [45]

    Song, PLOS ONE 16, 1 (2021)

    G. Song, PLOS ONE 16, 1 (2021)

    2021

  38. [46]

    H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H. Weissig, I. N. Shindyalov, and P. E. Bourne, Nucleic Acids Research 28, 235 (2000), https://academic.oup.com/nar/article- pdf/28/1/235/9895144/280235.pdf

    2000

  39. [47]

    Y. Chow, R. Pietranico, and A. Mukerji, Biochemical and Biophysical Research Communications 66, 1424 (1975)

    1975

  40. [48]

    R. E. Alcantara, C. Xu, T. G. Spiro, and V. Guallar, Proceedings of the National Academy of Sciences 104, 18451 (2007), https://www.pnas.org/doi/pdf/10.1073/pnas.0706026104. Supporting Information for1 Cooperative Function with Thermal Fluctuations in Mechanical Networks2 Ben ...

    2007 doi

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.