Pith. sign in

REVIEW 4 major objections 3 minor 75 references

SheafIQ: Sheaf-Theoretic Information Quantification of Vector Fields on Geometric Graphs

T0 review · 4 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper claims that the organization of vector fields on geometric graphs reduces to the normalized Shannon entropy of a sheaf residual energy field, and that this single scale-invariant descriptor exposes structure in proteins, brain…

desk verdict The sheaf machinery is decorative: the statistic actually used in applications is not a sheaf residual, and the sheaf residual itself reduces to Euclidean distance. read the letter →

arxiv 2608.08728 v1 pith:UB35ONIK submitted 2026-08-09 cs.IT math.IT

Cong Shen , Guancen Lin , Chuan-Shen Hu This is my paper · ORCID
classification cs.ITmath.IT
keywords cellularsheafvectorfieldongraphentropyresidualenergygeometricnetworkorganizationShannoninformationquantification
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 proposes SheafIQ, a method that turns the pattern of local mismatches between neighboring vectors on a graph into a single normalized entropy between 0 and 1. The idea is to express each node's vector in the coordinate frame of each incident edge, splitting it into parts parallel and perpendicular to the edge, so that disagreements between neighbors are measured relative to the geometry of the connection rather than in a global coordinate system. The size of each disagreement becomes an edge energy; after normalizing those energies to a probability distribution, ordinary Shannon entropy divided by the log of the number of edges summarizes how concentrated or spread out the local incompatibilities are. The paper argues that this one number captures organizational information about the vector field that graph topology alone and signal magnitude alone do not, and it demonstrates the claim on protein motions, Alzheimer's brain networks, urban traffic, and power grids.

What carries the argument

The load-bearing object is a cellular sheaf on the geometric graph whose edge stalks are $\mathbb{R}^{2d}$ and whose restriction maps are the edge-induced coordinate transformations of Eq. (1): a node vector is sent to its signed projection onto the edge direction together with its orthogonal complement, where $u_{ij}$ is the unit direction of the edge. This map puts both endpoint vectors into a common frame, so the sheaf residual measures genuine local incompatibility rather than an ordinary coordinate difference. From the residual norms an edge energy field $\varepsilon_{ij}=r_{ij}^2$ is built, normalized to a probability measure over edges, and the final descriptor is the Shannon entropy of that measure divided by $\log|E|$. The machinery's work is to separate magnitude, captured by total residual energy, from organization, captured by the relative distribution, and to make the comparison scale-invariant and graph-size-invariant.

What would settle it

On a fixed geometric graph with $|E|>1$, engineer two vector fields with identical total residual energy but with one field's residual energy placed on a single edge and the other spread uniformly; the framework predicts SheafIQ values of 0 and 1 respectively, so a failure to approach these bounds would falsify the entropy computation. For the complementary-information claim, regress $\hat{H}$ on total residual energy, mean signal magnitude, and a spectral graph entropy across many random vector fields on one graph: residuals near zero would show SheafIQ carries no information beyond the descriptors it claims to complement.

Watch

Extended reading notes

Core claim

The central claim is that a vector field on a geometrically embedded graph has a well-defined organizational signature: the normalized entropy of its sheaf residual energy distribution. For each edge, the two endpoint vectors are projected onto the edge direction and its orthogonal complement inside the edge stalk $\mathbb{R}^{2d}$; the difference of those projections is the sheaf residual. Squaring a stable scalar version of that residual gives an edge energy, the energies are normalized to a probability measure, and $\hat{H}=H/\log|E|$ is the SheafIQ score. The paper proves scale invariance under uniform rescaling of all energies and establishes the bounds $0\le \hat{H}\le 1$, with 0 meaning all incompatibility is concentrated on one edge and 1 meaning it is spread uniformly. Across the four application domains, the paper reports that this score and its associated residual hotspots reveal structure that conventional graph and signal descriptors miss.

Load-bearing premise

The load-bearing premise is that every node vector lives in the same Euclidean space as the geometric embedding ($q=d$), so the edge direction $u_{ij}$ defines a valid frame for the parallel/orthogonal decomposition; in the brain application this premise is not demonstrated, and if it fails the entire residual-energy construction is undefined.

Editorial extensions

If this is right

  • Vector fields on identical graphs can now be distinguished: where conventional graph entropy returns the same value for the same topology, SheafIQ changes with the organization of the node vectors.
  • Because the score is scale-invariant and normalized by $\log|E|$, organizations can be compared across graphs and datasets of different sizes on a common [0,1] scale.
  • Low SheafIQ localizes incompatibility: in the applications it flags mutation hotspots, AD-associated brain regions, congested traffic states, and critical power-grid buses and lines, while high SheafIQ indicates broadly distributed residual energy.
  • The response to localized electrical faults is monotone in severity and nearly absent for coherent global changes, suggesting the score isolates truly local organizational disruptions.
  • Edge-wise residual energies remain available, so SheafIQ is not just a scalar summary: it also produces interpretable hotspot maps and communication backbone structures.

Reading between the lines

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

  • Shannon entropy is permutation-invariant over edges, so SheafIQ cannot distinguish a clustered arrangement of high-residual edges from a scattered arrangement with the same histogram; a spatial variant weighted by graph distance would be a direct testable extension of the paper's own 'organization' language.
  • In the brain application (Section 3.3), the node vectors are never defined and q is never stated; if the node objects are time series or connectivity profiles rather than vectors in the same $\mathbb{R}^d$ as the AAL coordinates, the residual, energy, and entropy in that section are undefined, and the claim there should be read as conditional on the $q=d$ premise.
  • The paper uses $r_{ij}$, which suppresses directional variation of the perpendicular components, rather than the full residual norm $\|t_{ij}\|$; comparing both versions on the same datasets would reveal how much rotational information the current score discards.
  • The parallel/orthogonal split suggests a natural decomposition into gradient-like and curl-like residual energies; one could define separate 'parallel entropy' and 'orthogonal entropy' to characterize flow versus rotational organization on the graph.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The paper introduces SheafIQ, a descriptor intended to quantify the global organization of vector-valued states on geometrically embedded graphs. The construction proceeds by attaching a cellular sheaf to the graph, using edge-induced restriction maps to represent neighboring node vectors in a common edge stalk, defining a sheaf residual per edge, converting residual norms into edge energies, and finally taking the Shannon entropy of the normalized edge-energy distribution as the SheafIQ value. Propositions 1 and 2 establish standard entropy bounds and scale invariance for this distribution. The paper reports applications to protein dynamics, Alzheimer's disease functional networks, urban traffic data, and power grids, with claims that SheafIQ captures organizational information beyond topology-only and magnitude-only descriptors, supported by negative controls, baseline regressions, and multi-dataset replication.

Significance. If the sheaf-theoretic interpretation were valid, the paper would offer a general, parameter-free descriptor applicable to many scientific domains; the extensive empirical sections, including randomization controls in proteins, traffic regression residuals, and power-grid fault experiments, are a genuine strength and demonstrate care in validation. However, the significance as stated is not supported: the central sheaf residual reduces to the ordinary edge difference, the descriptor actually used in all applications is an ad hoc nonlinear statistic that is not a sheaf coboundary, and the brain application does not specify the vector field at all. Proposition 1 and Proposition 2 are correct, but they hold for any probability distribution over edges and therefore do not depend on the sheaf construction. As a result, the claimed novelty of a unified sheaf-theoretic information measure collapses, and the empirical findings must be reinterpreted as properties of a heuristic edge statistic, not of sheaf-theoretic incompatibility.

major comments (4)
  1. [Section 2.2.2, Eq. (1); Section 2.3.1, Eq. (3); Section 2.3.2] The sheaf residual t_ij is mathematically identical to the ordinary vector difference s_j - s_i. Substituting Eq. (1) into Eq. (3) gives the first block ((s_j - s_i)^T u_ij)u_ij and the second block (s_j - s_i) - ((s_j - s_i)^T u_ij)u_ij, so t_ij = s_j - s_i and ||t_ij|| = ||s_j - s_i||. The edge direction u_ij cancels completely. Consequently, the claimed 'edge-induced coordinate system' does not provide any geometric comparison beyond the Euclidean difference already used in graph signal processing, and the statement in Section 2.3.1 that the residual 'quantifies the local incompatibility ... relative to the underlying graph geometry' is unsupported. The sheaf construction is therefore equivalent to a trivial constant sheaf with identity-style restrictions, and the central theoretical novelty is not present.
  2. [Section 2.3.3, Eq. (5); Section 2.4; Section 2.5.2] The edge measure actually used in all downstream quantities — residual energy, the measure in Eq. (8), the probability in Eq. (9), the entropy in Eq. (10), and Definition 2 — is r_ij from Eq. (5), not the sheaf residual norm ||t_ij|| from Eq. (4). The statistic r_ij is not a sheaf coboundary and is not obtained from any cellular sheaf: it replaces the perpendicular vector difference by a difference of perpendicular magnitudes, which is a nonlinear operation. As a concrete counterexample, for u=(1,0), s_i=(0,1), and s_j=(0,-1), one has r_ij=0 while t_ij=(0,-2). Thus an edge can have zero residual energy under the application's statistic while being maximally incompatible under the stated sheaf definition. All application conclusions in Sections 3.1–3.5 therefore concern an ad hoc partial statistic, and the claims that SheafIQ quantifies 'sheaf incompatibility' or 'sheaf residual organization' are not justified. The authors must either re-derive every result for r_ij as an independent descriptor, with no appeal to sheaf theory, or replace r_ij by a genuinely sheaf-derived residual and rerun the applications.
  3. [Section 3.3] The brain-network application never specifies the node vectors or their dimension q, and no node coordinates or edge directions are defined. The SheafIQ framework requires q=d and an explicit geometric embedding with coordinates x_i to construct u_ij and the restriction maps of Eq. (1). If the node objects are regional time series, connectivity profiles, or any objects other than vectors in the same Euclidean space as the node coordinates, then the residual, energy, and entropy are not defined. The reader cannot verify the validity of any SheafIQ value reported in Section 3.3, and the central assumption of the method is unstated. This is a load-bearing omission because all subsequent hotspot and enrichment analyses inherit it.
  4. [Section 3.5] The power-grid application states that 'bus voltage magnitudes and phase angles obtained from AC power-flow analysis were treated as node signals (or node vectors)', but this does not specify a concrete vector construction. A phase angle is a scalar, and a magnitude-angle pair is not automatically a vector in R^d for the graph embedding; no node coordinates are defined for the IEEE 118-bus system. Without a precise specification of s_i, the embedding dimension d, and the edge directions, Eq. (1) and all derived quantities are undefined. The same ambiguity applies to the IEEE 300-bus generalization in Supplementary Section 6.3.
minor comments (3)
  1. [Section 2.5.2, proof of Proposition 1] In the proof, the symbol n is used for |E| without being defined; the statement should explicitly set n=|E| at the start.
  2. [Section 2.3.3, Eq. (5)] The sentence attributing the rotational invariance of r_ij to Eq. (2) is imprecise: Eq. (2) states equality of restriction maps under sign reversal of u_ij, while the claimed rigid-motion invariance of r_ij requires a separate argument and is not generally true if the parallel component changes under rotation or translation.
  3. [Section 3.1, Figure 2 caption] The caption for Figure 2(c) uses the phrase 'Normalized Information Quantity', which is not the term defined in Definition 2; the notation should be consistent with H_SheafIQ throughout.
Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The core descriptor has no fitted constants, but it depends on the q=d geometric assumption, on the ad hoc choice of r_ij as the incompatibility measure, and on the modeling decision to equate organization with Shannon entropy of normalized residual energies. The application results additionally depend on hand-chosen thresholds and on domain assumptions such as ANM modes representing protein dynamics.

free parameters (2)
  • Hotspot and pathway thresholds = top-5 residues, 10 Angstrom neighborhood, top-5 paths, Top-10/20 nodes
    Chosen by hand in each application; hotspot, pathway, and brain region results depend on these values, with only partial robustness checks.
  • Number of ANM modes = 20
    Section 3.1 restricts analysis to the first 20 non-rigid modes; reported mutant-specific SheafIQ changes are mode-dependent.
assumptions (5)
  • domain assumption Node vectors and geometric embedding have the same dimension (q=d), enabling parallel/orthogonal decomposition along each edge direction.
    Introduced in Section 2.1; used throughout and not verified for the brain dataset.
  • domain assumption The graph is embedded in R^d and u_ij=(x_j-x_i)/||x_j-x_i|| is a meaningful local reference frame.
    Section 2.2.2; if coordinates are arbitrary or noisy, the decomposition lacks physical meaning.
  • ad hoc to paper Shannon entropy of the normalized residual energy distribution is taken as the definition of global organizational complexity.
    Definition 2; this is the paper's modeling choice, not derived from an optimality criterion.
  • ad hoc to paper The ad hoc residual measure r_ij of Eq. (5), which suppresses perpendicular directional differences, is an acceptable edge incompatibility measure.
    All applications use r_ij; the claimed rigid-motion invariance in Section 2.3.3 is not established and appears false for translations.
  • domain assumption ANM normal modes adequately represent collective protein dynamics relevant to mutations and conformational change.
    Used in Sections 3.1 and 3.2 to generate node displacement vectors.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SheafIQ: Sheaf-Theoretic Information Quantification of Vector Fields on Geometric Graphs." pith.science (2026). https://pith.science/paper/UB35ONIK

@misc{pith2026260808728,
  author       = {Pith},
  title        = {Pith review of: SheafIQ: Sheaf-Theoretic Information Quantification of Vector Fields on Geometric Graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UB35ONIK}},
  note         = {Machine review of arXiv:2608.08728}
}
read the original abstract

Vector fields on graph structures naturally arise in diverse biological and engineered systems, where vector-valued states are defined on the nodes and evolve through the network interactions. Existing methods primarily characterize either the graph topology or individual signals, but generally do not quantify how local interactions among node-associated vectors are organized across the graph. To address this limitation, a sheaf-theoretic framework, termed SheafIQ, is proposed to represent neighboring vectors in a common edge-associated coordinate system, map local incompatibilities to a residual energy distribution, and quantify its global organization through entropy. Across proteins, functional brain networks, urban traffic systems, and power grids, SheafIQ consistently reveals complementary organizational information beyond conventional graph- and signal-based descriptors. More broadly, it establishes a unified information-theoretic framework for quantifying the organization of vector-valued states on geometric graphs, extending network analysis beyond graph topology alone.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

75 extracted references · 73 canonical work pages

  1. [1]

    S. H. Strogatz, Exploring complex networks.Nature410(6825), 268–276 (2001)

    work page 2001

  2. [2]

    Boguna, D

    M. Boguna, D. Krioukov, K. C. Claffy, Navigability of complex networks.Nature Physics5(1), 74–80 (2009)

    work page 2009

  3. [3]

    Ruths, D

    J. Ruths, D. Ruths, Control profiles of complex networks.Science343(6177), 1373–1376 (2014)

    work page 2014

  4. [4]

    Artime,et al., Robustness and resilience of complex networks.Nature Reviews Physics 6(2), 114–131 (2024)

    O. Artime,et al., Robustness and resilience of complex networks.Nature Reviews Physics 6(2), 114–131 (2024)

    work page 2024

  5. [5]

    Zheng, S

    W. Zheng, S. Doniach, A comparative study of motor-protein motions by using a simple elastic-network model.Proceedings of the National Academy of Sciences100(23), 13253– 13258 (2003)

    work page 2003

  6. [6]

    H.-J. Park, K. Friston, Structural and functional brain networks: from connections to cognition. Science342(6158), 1238411 (2013)

    work page 2013

  7. [7]

    D. S. Bassett, O. Sporns, Network neuroscience.Nature neuroscience20(3), 353–364 (2017)

    work page 2017

  8. [8]

    Herty, A

    M. Herty, A. Klar, Modeling, simulation, and optimization of traffic flow networks.SIAM Journal on Scientific Computing25(3), 1066–1087 (2003)

    work page 2003

Show all 75 references
  1. [9]

    G. A. Pagani, M. Aiello, The power grid as a complex network: a survey.Physica A: Statistical Mechanics and its Applications392(11), 2688–2700 (2013)

    2013

  2. [10]

    Donetti, P

    L. Donetti, P. I. Hurtado, M. A. Munoz, Entangled networks, synchronization, and optimal network topology.Physical Review Letters95(18), 188701 (2005)

    2005

  3. [11]

    C. Shen, X. Liu, J. Luo, K. Xia, Torsion graph neural networks.IEEE Transactions on Pattern Analysis and Machine Intelligence47(4), 2946–2956 (2025)

    2025

  4. [12]

    Dehmer, A

    M. Dehmer, A. Mowshowitz, A history of graph entropy measures.Information Sciences 181(1), 57–78 (2011). 33

    2011

  5. [13]

    N. Alon, A. Orlitsky, Source coding and graph entropies.IEEE Transactions on Information Theory42(5), 1329–1339 (1996)

    1996

  6. [14]

    C. Shen, G. Lin,𝜒-Entropy: A Distributional Measure of Multiscale Topological Organization. IEEE Transactions on Industrial Informatics(2026)

    2026

  7. [15]

    Sandryhaila, J

    A. Sandryhaila, J. M. Moura, Discrete signal processing on graphs.IEEE transactions on signal processing61(7), 1644–1656 (2013)

    2013

  8. [16]

    Sandryhaila, J

    A. Sandryhaila, J. M. Moura, Discrete signal processing on graphs: Frequency analysis.IEEE Transactions on signal processing62(12), 3042–3054 (2014)

    2014

  9. [17]

    Motsch, E

    S. Motsch, E. Tadmor, Heterophilious dynamics enhances consensus.SIAM review56(4), 577–621 (2014)

    2014

  10. [18]

    L. Yang, G. Song, R. L. Jernigan, Protein elastic network models and the ranges of cooperativity. Proceedings of the National Academy of Sciences106(30), 12347–12352 (2009)

    2009

  11. [19]

    Aminov,Geometry of vector fields(CRC Press) (2000)

    Y. Aminov,Geometry of vector fields(CRC Press) (2000)

    2000

  12. [20]

    do Goes, M

    F. do Goes, M. Desbrun, Y. Tong, Vector field processing on triangle meshes, inSIGGRAPH Asia 2015 Courses, SA ’15 (Association for Computing Machinery, New York, NY, USA) (2015), doi:10.1145/2818143.2818167,https://doi.org/10.1145/2818143.2818167

    2015

  13. [21]

    Crane, Discrete differential geometry: An applied introduction.Notices of the AMS, Com- munication1153(2) (2018)

    K. Crane, Discrete differential geometry: An applied introduction.Notices of the AMS, Com- munication1153(2) (2018)

    2018

  14. [22]

    Qiu,et al., Mapping transcriptomic vector fields of single cells.Cell185(4), 690–711 (2022)

    X. Qiu,et al., Mapping transcriptomic vector fields of single cells.Cell185(4), 690–711 (2022)

    2022

  15. [23]

    W. Yao, B. Lin, B. D. Anderson, M. Cao, Guiding vector fields for following occluded paths. IEEE Transactions on Automatic Control67(8), 4091–4106 (2022)

    2022

  16. [24]

    J. M. Curry,Sheaves, Cosheaves and Applications, Ph.D. thesis, University of Pennsylvania (2014). 34

    2014

  17. [25]

    J. M. Curry, Topological Data Analysis and Cosheaves.Japan Journal of Industrial and Applied Mathematics32(2), 333–371 (2015), doi:10.1007/s13160-015-0173-9,https://doi.org/ 10.1007/s13160-015-0173-9

    2015 doi

  18. [26]

    Curry, R

    J. Curry, R. Ghrist, V. Nanda, Discrete Morse Theory for Computing Cellular Sheaf Cohomology.Foundations of Computational Mathematics16(4), 875–897 (2016), doi: 10.1007/s10208-015-9266-8,https://doi.org/10.1007/s10208-015-9266-8

    2016 doi

  19. [27]

    Hansen, R

    J. Hansen, R. Ghrist, Toward a spectral theory of cellular sheaves.Journal of Applied and Computational Topology3(4), 315–358 (2019)

    2019

  20. [28]

    Battiloro, Z

    C. Battiloro, Z. Wang, H. Riess, P. Di Lorenzo, A. Ribeiro, Tangent bundle convolutional learning: from manifolds to cellular sheaves and back.IEEE Transactions on Signal Processing 72, 1892–1909 (2024)

    2024

  21. [29]

    Bodnar, F

    C. Bodnar, F. Di Giovanni, B. Chamberlain, P. Lio, M. Bronstein, Neural sheaf diffusion: A topological perspective on heterophily and oversmoothing in GNNs.Advances in Neural Information Processing Systems35, 18527–18541 (2022)

    2022

  22. [30]

    Cooperband, A

    Z. Cooperband, A. McRobie, C. Millar, B. Schulze, Homology of moment frames, inPro- ceedings of IASS Annual Symposia(International Association for Shell and Spatial Structures (IASS)) (2024), pp. 1–10

    2024

  23. [31]

    Cooperband, R

    Z. Cooperband, R. Ghrist, Towards Homological Methods in Graphic Statics.Journal of the International Association for Shell and Spatial Structures64(4), 266–277 (2023), doi: 10.20898/j.iass.2023.024

    2023 doi

  24. [32]

    Hansen, R

    J. Hansen, R. Ghrist, Learning sheaf laplacians from smooth signals, inICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)(IEEE) (2019), pp. 5446–5450

    2019

  25. [33]

    Robinson, Assignments to sheaves of pseudometric spaces.Compositionality2(2020)

    M. Robinson, Assignments to sheaves of pseudometric spaces.Compositionality2(2020)

    2020

  26. [34]

    A. M. Moustafa, M. B. Abdelghany, Finite-Trace Cellular Sheaves for Validation and Estimation of Hybrid Dynamical Systems.IEEE Control Systems Letters(2026). 35

    2026

  27. [35]

    Hu, A Sheaf-Theoretic and Topological Perspective on Complex Network Modeling and Attention Mechanisms in Graph Neural Models (2026),https://arxiv.org/abs/2601

    C.-S. Hu, A Sheaf-Theoretic and Topological Perspective on Complex Network Modeling and Attention Mechanisms in Graph Neural Models (2026),https://arxiv.org/abs/2601. 21207

    2026

  28. [36]

    Zhao,et al., Asynchronous Nonlinear Sheaf Diffusion for Multi-Agent Coordination (2025), https://arxiv.org/abs/2510.00270

    Y. Zhao,et al., Asynchronous Nonlinear Sheaf Diffusion for Multi-Agent Coordination (2025), https://arxiv.org/abs/2510.00270

    2025

  29. [37]

    Martinez Esturo, C

    J. Martinez Esturo, C. R ¨ossl, H. Theisel, Smoothed quadratic energies on meshes.ACM Transactions on Graphics (TOG)34(1), 1–12 (2014)

    2014

  30. [38]

    Sloboda, Sheaf-Laplacian Obstruction and Projection Hardness for Cross-Modal Compati- bility on a Modality-Independent Site.arXiv preprint arXiv:2604.07632(2026)

    T. Sloboda, Sheaf-Laplacian Obstruction and Projection Hardness for Cross-Modal Compati- bility on a Modality-Independent Site.arXiv preprint arXiv:2604.07632(2026)

    2026 arXiv

  31. [39]

    Osting, C

    B. Osting, C. D. White, ´E. Oudet, Minimal Dirichlet energy partitions for graphs.SIAM Journal on Scientific Computing36(4), A1635–A1651 (2014)

    2014

  32. [40]

    Y. Ji,et al., How Does Topology Bias Distort Message Passing in Graph Recommender? A Dirichlet Energy Perspective.Advances in Neural Information Processing Systems38, 45430– 45458 (2026)

    2026

  33. [41]

    F. Chen, G. Cheung, X. Zhang, Manifold graph signal restoration using gradient graph Lapla- cian regularizer.IEEE Transactions on Signal Processing72, 744–761 (2024)

    2024

  34. [42]

    Bromiley, N

    P. Bromiley, N. Thacker, E. Bouhova-Thacker, Shannon entropy, Renyi entropy, and informa- tion.Statistics and Inf. Series (2004-004)9(2004), 2–8 (2004)

    2004

  35. [43]

    C. A. Smith,et al., Allosteric switch regulates protein–protein binding through collective motion.Proceedings of the National Academy of Sciences113(12), 3269–3274 (2016)

    2016

  36. [44]

    H. J. Berendsen, S. Hayward, Collective protein dynamics in relation to function.Current opinion in structural biology10(2), 165–169 (2000)

    2000

  37. [45]

    Liu,et al., Dynamical transition of collective motions in dry proteins.Physical Review Letters119(4), 048101 (2017)

    Z. Liu,et al., Dynamical transition of collective motions in dry proteins.Physical Review Letters119(4), 048101 (2017)

    2017

  38. [46]

    S. M. Kelly, Anisotropic networks.Journal of Materials Chemistry5(12), 2047–2061 (1995). 36

    1995

  39. [47]

    Eyal, L.-W

    E. Eyal, L.-W. Yang, I. Bahar, Anisotropic network model: systematic evaluation and a new web interface.Bioinformatics22(21), 2619–2627 (2006)

    2006

  40. [48]

    A. R. Atilgan,et al., Anisotropy of fluctuation dynamics of proteins with an elastic network model.Biophysical journal80(1), 505–515 (2001)

    2001

  41. [49]

    B. W. Matthews, Studies on protein stability with T4 lysozyme.Advances in protein chemistry 46, 249–278 (1995)

    1995

  42. [50]

    Bouvignies,et al., Solution structure of a minor and transiently formed state of a T4 lysozyme mutant.Nature477(7362), 111–114 (2011)

    G. Bouvignies,et al., Solution structure of a minor and transiently formed state of a T4 lysozyme mutant.Nature477(7362), 111–114 (2011)

    2011

  43. [51]

    Sideraki, W

    V. Sideraki, W. Huang, T. Palzkill, H. F. Gilbert, A secondary drug resistance mutation of TEM-1𝛽-lactamase that suppresses misfolding and aggregation.Proceedings of the National Academy of Sciences98(1), 283–288 (2001)

    2001

  44. [52]

    Jacquier,et al., Capturing the mutational landscape of the beta-lactamase TEM-1.Proceed- ings of the National Academy of Sciences110(32), 13067–13072 (2013)

    H. Jacquier,et al., Capturing the mutational landscape of the beta-lactamase TEM-1.Proceed- ings of the National Academy of Sciences110(32), 13067–13072 (2013)

    2013

  45. [53]

    J. L. Radkiewicz, C. L. Brooks, Protein dynamics in enzymatic catalysis: exploration of dihy- drofolate reductase.Journal of the American Chemical Society122(2), 225–231 (2000)

    2000

  46. [54]

    J. R. Schnell, H. J. Dyson, P. E. Wright, Structure, dynamics, and catalytic function of dihy- drofolate reductase.Annu. Rev. Biophys. Biomol. Struct.33(1), 119–140 (2004)

    2004

  47. [55]

    Lumry, H

    R. Lumry, H. Eyring, Conformation changes of proteins.The Journal of physical chemistry 58(2), 110–120 (1954)

    1954

  48. [56]

    D. E. Koshland, Conformational changes: how small is big enough?Nature medicine4(10), 1112–1114 (1998)

    1998

  49. [57]

    Arora, C

    K. Arora, C. L. Brooks III, Large-scale allosteric conformational transitions of adenylate kinase appear to involve a population-shift mechanism.Proceedings of the National Academy of Sciences104(47), 18496–18501 (2007)

    2007

  50. [58]

    Olsson, M

    U. Olsson, M. Wolf-Watz, Overlap between folding and functional energy landscapes for adenylate kinase conformational change.Nature communications1(1), 111 (2010). 37

    2010

  51. [59]

    C. Tang, C. D. Schwieters, G. M. Clore, Open-to-closed transition in apo maltose-binding protein observed by paramagnetic NMR.Nature449(7165), 1078–1082 (2007)

    2007

  52. [60]

    Y. Shen, A. Bax, Two-conformer equilibrium of maltose-binding protein in the absence of ligand from residual dipolar coupling analysis.Protein Science35(1), e70425 (2026)

    2026

  53. [61]

    W. S. Bennett Jr, T. A. Steitz, Glucose-induced conformational change in yeast hexokinase. Proceedings of the National Academy of Sciences75(10), 4848–4852 (1978)

    1978

  54. [62]

    C. J. Stam,et al., Graph theoretical analysis of magnetoencephalographic functional connec- tivity in Alzheimer’s disease.Brain132(1), 213–224 (2009)

    2009

  55. [63]

    C. J. Stam, B. Jones, G. Nolte, M. Breakspear, P. Scheltens, Small-world networks and func- tional connectivity in Alzheimer’s disease.Cerebral cortex17(1), 92–99 (2007)

    2007

  56. [64]

    R. C. Petersen,et al., Alzheimer’s disease Neuroimaging Initiative (ADNI) clinical character- ization.Neurology74(3), 201–209 (2010)

    2010

  57. [65]

    N. Tzourio-Mazoyer,et al., Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain.Neuroimage15(1), 273–289 (2002)

    2002

  58. [66]

    M. D. Greicius, G. Srivastava, A. L. Reiss, V. Menon, Default-mode network activity distin- guishes Alzheimer’s disease from healthy aging: evidence from functional MRI.Proceedings of the National Academy of Sciences101(13), 4637–4642 (2004)

    2004

  59. [67]

    Y. L. Rao,et al., Hippocampus and its involvement in Alzheimer’s disease: a review.3 Biotech 12(2), 55 (2022)

    2022

  60. [68]

    C. J. Galton,et al., Differing patterns of temporal atrophy in Alzheimer’s disease and semantic dementia.Neurology57(2), 216–225 (2001)

    2001

  61. [69]

    G. J. Upton, Fisher’s exact test.Journal of the Royal Statistical Society: Series A (Statistics in Society)155(3), 395–402 (1992)

    1992

  62. [70]

    C.-F. J. Wu, Jackknife, bootstrap and other resampling methods in regression analysis.the Annals of Statistics14(4), 1261–1295 (1986). 38

    1986

  63. [71]

    Hesterberg, Bootstrap.Wiley Interdisciplinary Reviews: Computational Statistics3(6), 497–526 (2011)

    T. Hesterberg, Bootstrap.Wiley Interdisciplinary Reviews: Computational Statistics3(6), 497–526 (2011)

    2011

  64. [72]

    Y. Li, R. Yu, C. Shahabi, Y. Liu, Diffusion Convolutional Recurrent Neural Network: Data- Driven Traffic Forecasting, in6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings (OpenReview....

    2018

  65. [73]

    Dorfman, A formula for the Gini coefficient.The review of economics and statisticspp

    R. Dorfman, A formula for the Gini coefficient.The review of economics and statisticspp. 146–149 (1979)

    1979

  66. [74]

    L. Thurner,et al., pandapower—an open-source python tool for convenient modeling, analysis, and optimization of electric power systems.IEEE Transactions on Power Systems33(6), 6510–6521 (2018)

    2018

  67. [75]

    Khanabadi, H

    M. Khanabadi, H. Ghasemi, M. Doostizadeh, Optimal transmission switching considering voltage security and N-1 contingency analysis.IEEE Transactions on Power Systems28(1), 542–550 (2012). 39 Supplementary Materials for SheafIQ: Sheaf-Theoretic Information Quantification of Vec...

    2012

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.