Pith. sign in

REVIEW 4 major objections 4 minor 47 references

Physics-Based Molecular Fingerprints from Spectral Graph Theory Provide Efficient Geometry-Aware Measures of Chemical Similarity

T0 review · 4 major / 4 minor · reviewed 2026-08-08 · 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 graph-Laplacian eigenvalues of a complete graph over 3D atomic coordinates form a fixed-length, permutation-invariant fingerprint that distinguishes conformers and stereoisomers that 2D connectivity collapses, at a…

desk verdict A genuinely useful new 3D fingerprint that distinguishes conformers and stereoisomers cheaply, but its generalizability claim rests on a dataset-specific M=16 truncation and the clustering benchmarks need tightening. read the letter →

arxiv 2608.05336 v1 pith:QPPDVTJR submitted 2026-08-05 physics.chem-ph cs.LG

Jacob W. Toney , Ayleen Y. Farnood , Samir Darouich , Heather J. Kulik This is my paper · ORCID
classification physics.chem-phcs.LG
keywords spectralgraphtheorymolecularfingerprintchemicalsimilarityLaplacian3Drepresentationstereoisomerdiscriminationcheminformaticscommunitydetection
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

The paper introduces a molecular fingerprint built from the eigenvalues of graph Laplacians on a complete graph whose edge weights are four physics-inspired pairwise interactions derived from 3D coordinates and tabulated atomic properties. The authors aim to show that this fixed-length fingerprint is geometry-aware without the cost of pairwise alignment, and that it distinguishes structures such as conformers, stereoisomers, and mutated proteins that 2D connectivity representations collapse together. If correct, the method would give cheminformatics an interpretable, cheap, and domain-general similarity measure for screening large chemical spaces, with immediate uses in clustering, nearest-neighbor property estimation, and applicability-domain analysis.

What carries the argument

The central object is the multichannel graph Laplacian of a complete graph whose vertices are atoms. For each of four physical channels, the Laplacian is constructed from a distance-weighted adjacency matrix, and its eigenvalues are collected; the smallest nonzero eigenvalue (the algebraic connectivity) is placed first and the remaining extreme eigenvalues follow, truncated to 16 per channel after Frobenius normalization of the adjacency matrices. Because eigenvalues are invariant to vertex relabeling and the interaction weights depend only on interatomic distances, the resulting 64-component vector is permutation-invariant and E(3)-invariant, and it is fixed-length regardless of molecular size.

What would settle it

A direct falsifier would be a pair of stereoisomers or conformers with identical 2D connectivity, known different properties, and nearly identical spectral fingerprints (similarity above about 0.99), while unrelated molecules score lower; that would show the truncation discards the discriminating geometry. A quantitative version is to scan a large conformational ensemble and check whether pairwise spectral distances order the molecules the same way as RMSD or energy differences; any systematic inversion would contradict the claimed locality.

Watch

Extended reading notes

Core claim

The paper's central claim is that the spectrum of a graph Laplacian built from a complete graph over 3D atomic coordinates encodes molecular similarity in a way that is simultaneously geometry-aware, permutation-invariant, alignment-free, and fixed-length. The fingerprint uses four channels—electrostatics, bonding, sterics, and dispersion—whose edge weights are heuristic physical functions of interatomic distance and tabulated atomic properties; the channel-wise graph Laplacians are formed as $L^{(k)} = D^{(k)} - A^{(k)}$, and for each channel the smallest nonzero eigenvalue is placed first and the largest eigenvalues are kept down to a length of 16, giving a 64-dimensional vector. The authors report that this vector separates chair from boat cyclohexane, cis from trans platinum complexes, fac from mer iridium isomers, and proteins differing by side-chain mutations, all cases where 2D connectivity fingerprints coincide. They further report that clusters formed from fingerprint neighborhoods track property distributions across organic, inorganic, biological, reaction, and framework datasets, and that nearest-neighbor predictions in fingerprint space rank-correlate strongly with several quantum-chemical properties.

Load-bearing premise

The load-bearing premise is that a fixed 16-eigenvalue truncation, combined with Frobenius normalization, preserves enough chemical information to compare molecules that differ greatly in size; the paper states this cutoff was chosen empirically and that results for large proteins depend on it.

Editorial extensions

If this is right

  • Molecule libraries can be screened without computing pairwise alignments, because similarity is just Euclidean distance between fixed-length vectors.
  • Stereoisomers and conformers that are invisible to 2D fingerprints become distinguishable from coordinates alone.
  • The fingerprint gives a training-free property estimate: nearby molecules in fingerprint space have similar size-extensive properties such as atomization energy and polarizability.
  • It offers a model-agnostic applicability-domain estimate that can be computed before any model is trained.
  • Reactions can be compared by concatenated reactant and product fingerprints, and reaction similarity tracks activation-energy differences without transition-state geometries.

Reading between the lines

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

  • A natural next test is whether the default 16-eigenvalue cutoff can be replaced by an adaptive per-molecule cutoff chosen from the eigenvalue decay, which would remove the main tuned hyperparameter.
  • Because the fingerprint needs only 3D coordinates, it could serve as a cheap geometry-aware uncertainty signal for machine-learned property models even when those models are trained on 2D inputs.
  • The same complete-graph construction could extend to periodic solids or supramolecular assemblies by using cutoffs and supercell coordinates, provided the channel formulas are rescaled consistently.
  • If the reported locality is robust, spectral fingerprint distances could be used to design training sets that deliberately span underrepresented geometries rather than random splits.
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 introduces 'spectral fingerprints': for each molecule, four weighted complete graphs are constructed from 3D atomic coordinates using heuristic electrostatics, bonding, sterics, and dispersion interaction functions; the eigenvalues of each channel's graph Laplacian are truncated to a fixed length (M=16 per channel) and concatenated into a length-64 vector. Similarity is defined as an inverse Euclidean distance. The representation is argued to be permutation-invariant, E(3)-invariant, alignment-free, and cheap, and is demonstrated on conformers, stereoisomers, reactions, MOFs, and proteins. The authors benchmark Leiden-based clustering against scaffolds, Morgan, RACs, and CD-RACs on six datasets (QM9, tmQMg, QCell, Transition1x, QMOF, GEMS), and evaluate k-NN regression and applicability-domain estimation for QM9 and tmQMg. The central claim is that spectral fingerprints provide a generalizable, interpretable, geometry-aware similarity measure that overcomes 2D limitations at low computational cost.

Significance. If the claims hold, this is a practically useful contribution: a fixed-length, alignment-free 3D descriptor with an interpretable construction and no fitted target-property model. The paper ships an open-source implementation, a Zenodo repository, and machine-checked case studies; the representation is not trained on target properties, and the evaluation is largely external, which is a strength. The clearest value is in settings where 2D connectivity is insufficient (stereoisomers, conformers, coordination geometry) and where alignment-based 3D methods are too costly. However, the main benchmarking and the large-system demonstrations rest on empirically tuned hyperparameters and on comparisons across different parsable subsets and cluster counts, so the strength of the generalizability claim is currently not fully supported.

major comments (4)
  1. [§3a, Eq. (8); SI Figures S1–S2] The default truncation M=16 is a tuned hyperparameter, not a derived guarantee. The text states that M 'was determined empirically' and that 'larger chemical systems beyond the scope of small molecules would benefit from dataset-specific tuning'; SI Figure S2 explicitly shows that protein similarities depend on the number of eigenvalues, while the plateau used to select M=16 (Figure S1) was established on small systems. Because the abstract and §3b claim generalizability to macromolecules and reticular chemistry, the reported protein similarities (0.976 and 0.436) and the QMOF/GEMS clustering results are conditional on an M value that may be suboptimal for large systems. Please provide an M-robustness analysis: report the P450/WelO5 similarities and the key Table 1–2 metrics for M = 8, 16, 32, 64, and show that method rankings are stable, or explicitly temper the generalizability claim to the small-molecule regime.
  2. [Tables 1 and 2; §3c] The benchmark comparisons are made across different parsable subsets and different numbers of clusters, which undermines the 'best overall performance' conclusion. For example, on QM9 the number of clusters is 76 for spectral, 16 for Morgan, and 1,132 for scaffolds; on QCell, scaffold parsing succeeds for only 0.4% of the dataset. Since the U-test percentage, DBI, and CHI depend on both the population and the number of clusters, the values in Tables 1–2 are not directly comparable across representations. The k-NN analysis already includes a parsability-controlled comparison (Tables S12–S14); please perform the analogous common-subset evaluation for the clustering benchmarks and report cluster counts alongside all metrics.
  3. [§3c, Tables 1–2; Text S3] The primary extrinsic metric is the percentage of cluster pairs with statistically significant Mann–Whitney U tests at α=0.05, reported without any multiple-testing correction or confidence interval. With 76 clusters there are 2,850 pairwise tests, and at α=0.05 one expects roughly 5% false positives even under the null; the expected number of significant pairs grows quadratically with the number of clusters. This makes the reported percentages (e.g., 98.0% for QM9 spectral, 87.3% for tmQMg) hard to interpret across methods that produce very different numbers of clusters. Please report adjusted p-values (e.g., Benjamini–Hochberg), or effect sizes with confidence intervals, or at least state the number of pairwise tests and the expected false-positive rate for each row.
  4. [§3d; SI Tables S9–S11] The k-NN results are reported for k=5, but the text states that k=5 'is observed to maximize the mean rank correlation across all tasks and representations' after evaluating k ∈ {1, 5, 25}. Selecting k on the same evaluation data can inflate the reported Spearman correlations and may favor the method for which the selected k is most beneficial. The SI does contain results for k=1 and k=25, which is helpful; please state explicitly that k was tuned on the evaluation set and show the sensitivity of the main conclusions to k (e.g., whether spectral fingerprints still achieve the highest rank correlations at k=1 and k=25 on the same tasks).
minor comments (4)
  1. [§3a, Eq. (4)] The bonding-channel formula is not rendered clearly; the denominator in the exponential appears to be missing in the typeset equation. Please define it explicitly, e.g., as exp(−r_ij / (r_cov,i + r_cov,j)).
  2. [Tables 1 and 2] The table headers state 'Bold indicates the best result for a given metric, and the arrow indicates what direction corresponds to a better value,' but no arrows appear in the printed headers. Adding ↑/↓ to the metric names would make the tables self-explanatory.
  3. [§3c, Supporting Information Table S8] The text says RACs vectors 'consistently require over ten times longer' than spectral fingerprints; for QCell the ratio is over 40×. The statement is technically true but understates the observed gap; consider giving the range of speedups.
  4. [§3b] In the paragraph on Transition1x reactions, the text reports similarities of 0.434 and 0.429 between the ring-opening reaction and the two oxidation/reduction reactions, but Figure 2c is described as showing 'reaction spectral fingerprints between (reactant, product) pairs'; please clarify in the figure caption whether the displayed values correspond to pairwise reaction similarities or to the concatenated fingerprints themselves.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the spectral fingerprint is not fitted to target properties and the main benchmarks are external; disclosed hyperparameter choices are limitations, not construction-level circularity.

full rationale

The derivation chain is self-contained. The fingerprint is constructed from 3D coordinates via four explicitly heuristic physics-inspired interaction channels (Eqs. 3–6), channel-wise graph Laplacians (Eq. 7), and truncation to M=16 eigenvalues per channel (Eq. 8); similarity is then defined as Euclidean distance (Eq. 9). None of these equations is defined in terms of the target properties (atomization energy, dipole, gap, polarizability, band gap) or the cluster labels used for evaluation. The main evaluations use external benchmarks (QM9, tmQMg, QCell, Transition1x, QMOF, GEMS) and compare against independent baselines, so the central claims do not reduce to a fit. The paper transparently states that M=16 was determined empirically and that the method was initially developed on systems that later appear as demonstrations (Section 2; SI Figures S1–S2), and that k=5 in k-NN regression was selected by maximizing mean rank correlation across the same tasks (Section 3d). These are hyperparameter-selection and in-sample-demonstration limitations, and the M-robustness caveat for large proteins is explicitly acknowledged in SI Figure S2, but they are not cases where a predicted quantity is identical to an input by construction. The authors' self-citations (molSimplify, RACs/CD-RACs, ElemeNet) are used as tools or baselines, not as load-bearing validation of the fingerprint. No circular step can therefore be exhibited.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The method leans on several hand-chosen hyperparameters and heuristic interaction forms. No new physical entities are introduced, but the usefulness of the fingerprint depends on the empirical validity of the channel equations, the M=16 truncation, and the normalization scheme.

free parameters (5)
  • M (eigenvalues per channel) = 16
    Default fingerprint length, chosen empirically from a range of 2 to 128; Supporting Information Figure S2 shows large systems depend on M.
  • Distance cutoff for large systems = 10 Å
    Sparsification cutoff for proteins with over 1,000 atoms; chosen for tractability and changes which pairwise interactions are included.
  • Similarity graph k = 5 log N
    Heuristic neighborhood size for Leiden clustering; the factor 5 is chosen without systematic tuning.
  • Leiden resolution parameter = 1
    Resolution in the Reichardt-Bornholdt Potts model; affects the number of clusters produced.
  • k for k-NN regression = 5
    Selected among {1, 5, 25} as maximizing mean rank correlation on the evaluation tasks; reported results use this tuned value.
assumptions (5)
  • standard math Eigenvalues of a graph Laplacian are permutation invariant and encode global and local graph structure.
    Standard spectral graph theory, cited via Chung 1997 and used throughout Section 3a.
  • domain assumption The four heuristic channel forms in Eqs. 3-6 capture chemically relevant interactions without fitted coefficients.
    Coulombic, Slater-type bonding, steric overlap, and London dispersion are physically motivated but heuristic; their relative balance is not derived from data or theory.
  • ad hoc to paper Frobenius normalization and distance normalization by covalent radii make molecules of different sizes and compositions comparable.
    Introduced without a uniqueness or invariance argument; the paper notes degeneracy for N=2 and size-dependent behavior in Figure S2.
  • ad hoc to paper Truncating each channel to M=16 eigenvalues preserves enough information for chemical similarity.
    M is chosen empirically from 2 to 128, and Supporting Information Figure S2 shows results depend on M for large chemical systems.
  • domain assumption The similar property principle holds for the test datasets and target properties used in clustering and k-NN evaluation.
    The evaluation assumes that molecules nearby in representation space have similar properties, which is the standard cheminformatics premise invoked in the introduction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics-Based Molecular Fingerprints from Spectral Graph Theory Provide Efficient Geometry-Aware Measures of Chemical Similarity." pith.science (2026). https://pith.science/paper/QPPDVTJR

@misc{pith2026260805336,
  author       = {Pith},
  title        = {Pith review of: Physics-Based Molecular Fingerprints from Spectral Graph Theory Provide Efficient Geometry-Aware Measures of Chemical Similarity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QPPDVTJR}},
  note         = {Machine review of arXiv:2608.05336}
}
read the original abstract

Molecular representations are essential for the evaluation of molecular similarity and the development of structure-property relationships. Despite the known importance of 3D structure to determine chemical and physical properties, the most widely used molecular fingerprints encode only two-dimensional connectivity. Such representations fail to distinguish similar but distinct stereoisomers and conformers. Alternative 3D methods are typically defined pairwise, making their application to large chemical spaces prohibitive, while deep learning embeddings are expressive but uninterpretable and limited by their training data diversity. Here, we introduce novel physics-inspired molecular fingerprints based on principles from spectral graph theory. We represent molecules as a complete graph in 3D space, with edge weights encoding heuristic physical interactions. Eigenvalue decomposition of the resulting graph Laplacian matrix results in a computationally efficient fixed-length chemical fingerprint that encodes 3D structure while obeying necessary physical symmetries of permutation and E(3) invariance. Spectral fingerprints differentiate between unique molecular structures with identical 2D connectivity, overcoming a limitation of 2D descriptors, while maintaining the low computational cost needed for efficient screening of vast chemical spaces. We evaluate our fingerprints with community detection algorithms and observe strong performance against representative baselines across datasets from organic, inorganic, biological, reticular, and reaction chemistry. Nearest-neighbor property estimation and applicability domain analyses reveal the utility of our molecular representation in machine learning and cheminformatics. We anticipate that spectral fingerprints will serve as generalizable, interpretable, and efficient measures of chemical similarity that incorporate 3D information at minimal cost.

Figures

Figures reproduced from arXiv: 2608.05336 by the authors.

Figure 1
Figure 1. Illustrative procedure for generating spectral fingerprints from 3D atomic coordinates. a) 3D coordinates of glycine optimized with UFF. Functional forms for pairwise interactions between atom pairs (𝑖,𝑗) are shown for electrostatic, bonding, steric, and dispersion channels. Atoms are colored as follows: H in white, C in gray, N in blue, O in red. b) Heatmaps of normalized adjacency matrices for electrostatic, bondi… view at source ↗
Figure 2
Figure 2. Representative chemical systems analyzed using spectral fingerprints. a) Boat and chair conformers of cyclohexane (C6H12) exhibit a high spectral similarity of 0.961. b) cis and trans stereoisomers of Pt(II)Cl2(NH3)2 are described by distinct spectral fingerprints. c) Reaction spectral fingerprints between (reactant, product) pairs of oxidation and reduction reactions from the Transition1x dataset. d) in silico Gly-… view at source ↗
Figure 4
Figure 4. Results of k-NN regression analysis performed on QM9 and tmQMg datasets in molecular representation space. For a given dataset, target, and representation, Spearman rank correlation is reported between predicted and true property values for each molecule, where predicted values are calculated as the average between the 𝑘 = 5 nearest neighbors in representation space. Spectral, RACs, and CD-RACs are each evaluated on… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 41 canonical work pages

  1. [1]

    J Chem Inf Comput Sci 1988, 28, 31-36

    Introduction to Methodology and Encoding Rules. J Chem Inf Comput Sci 1988, 28, 31-36. (2) Wigh, D. S.; Goodman, J. M.; Lapkin, A. A. A Review of Molecular Representation in the Age of Machine Learning. WIREs Comput Mol Sci 2022,

    work page 1988

  2. [2]

    M.; Downs, G

    (41) Willett, P.; Barnard, J. M.; Downs, G. M. Chemical Similarity Searching. J Chem Inf Comput Sci 1998, 38, 983-996. (42) Barbosa, F.; Horvath, D. Molecular Similarity and Property Similarity. Current Topics in Medicinal Chemistry 2004, 4, 589-600. 39 (43) Nikolova, N.; Jaworska, J. Approaches to Measure Chemical Similarity – a Review. Mol Inform 2004, ...

    work page 1998

  3. [3]

    (24) Rupp, M.; Tkatchenko, A.; Müller, K.-R.; Lilienfeld, O. A. v. Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning. Phys Rev Lett 2012,

    work page 2012

  4. [4]

    Conclusions In summary, we developed physics-based molecular fingerprints to encode 3D chemical structure using principles from spectral graph theory. To address the gap between fixed-length 2D representations and typically expensive 3D-aware methods, we encoded each molecule as a weighted complete graph over its 3D atomic coordinates, assigning edge weig...

  5. [5]

    J Med Chem 1996, 39, 2887-2893

    Molecular Frameworks. J Med Chem 1996, 39, 2887-2893. (6) Rogers, D.; Hahn, M. Extended-Connectivity Fingerprints. J Chem Inf Model 2010, 50, 742-754. (7) Gobbi, A.; Poppinger, D. Genetic Optimization of Combinatorial Libraries. Biotechnol Bioeng 1998, 61, 47-54. 37 (8) Nandy, A.; Duan, C.; Taylor, M. G.; Liu, F.; Steeves, A. H.; Kulik, H. J. Computationa...

    work page 1996

  6. [7]

    (14) Hansen, K.; Biegler, F.; Ramakrishnan, R.; Pronobis, W.; Lilienfeld, O. A. v.; Müller, K.-R.; Tkatchenko, A. Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space. J Phys Chem Lett 2015, 6, 2326-2331. (15) Janet, J. P.; Kulik, H. J. Resolving Transition Metal Chemical Space: Feature Sele...

    work page 2015

  7. [8]

    G.; Heras-Domingo, J.; Kitchin, J

    (91) Garrison, A. G.; Heras-Domingo, J.; Kitchin, J. R.; Dos Passos Gomes, G.; Ulissi, Z. W.; Blau, S. M. Applying Large Graph Neural Networks to Predict Transition Metal Complex Energies Using the Tmqm_Wb97mv Data Set. J Chem Inf Model 2023, 63, 7642-7654. (92) Ioannidis, E. I.; Gani, T. Z. H.; Kulik, H. J. Molsimplify: A Toolkit for Automating Discovery...

    work page 2023

  8. [9]

    W.; Farnood, A.; Darouich, S.; Kulik, H

    (99) Toney, J. W.; Farnood, A.; Darouich, S.; Kulik, H. J., 2026, DOI:10.5281/zenodo.21614395 10.5281/zenodo.21614395. (100) Mann, H. B.; Whitney, D. R. On a Test of Whether One of Two Random Variables Stochastically Larger Than the Other. Ann Math Stat 1947, 18, 50-60. (101) Virtanen, P.; Gommers, R.; Oliphant, T. E.; Haberland, M.; Reddy, T.; Cournapeau...

    work page doi:10.5281/zenodo.21614395 2026

Show all 47 references
  1. [10]

    Improving the Reliability of Molecular String Representations for Generative Chemistry

    (11) Reboul, E.; Wefers, Z.; Prabakaran, H.; Waldispühl, J.; Taly, A. Improving the Reliability of Molecular String Representations for Generative Chemistry. J Chem Inf Model 2025, 65, 10221-10238. (12) Rasmussen, M. H.; Strandgaard, M.; Seumer, J.; Hemmingsen, L. K.; Frei, A....

    2025

  2. [12]

    A.; Vassilev-Galindo, V.; Cheng, B.; Chmiela, S.; Gastegger, M.; Müller, K.-R.; Tkatchenko, A

    (3) Keith, J. A.; Vassilev-Galindo, V.; Cheng, B.; Chmiela, S.; Gastegger, M.; Müller, K.-R.; Tkatchenko, A. Combining Machine Learning and Computational Chemistry for Predictive Insights into Chemical Systems. Chem Rev 2021, 121, 9816-9872. (4) Butler, K. T.; Davies, D. W.; C...

    2021

  3. [13]

    (36) Ding, T.; Larrea-Gallegos, G.; Busio, F.; Marvuglia, A.; Schaubroeck, T. Characterizing Chemical Toxicity for Life Cycle Assessment Using Machine Learning and Deep Learning Models Based on Environmental Footprint-Methodological Comparison & Textile Case Study. RSC Sustain...

    2026 doi

  4. [15]

    The Shark Integral Generation and Digestion System

    (83) Neese, F. The Shark Integral Generation and Digestion System. J Comp Chem 2022,

    2022

  5. [17]

    (13) Kulik, H. J. Making Machine Learning a Useful Tool in the Accelerated Discovery of Transition Metal Complexes. WIREs Comput Mol Sci 2019,

    2019

  6. [18]

    S.; Grant, J

    (46) Rush, T. S.; Grant, J. A.; Mosyak, L.; Nicholls, A. A Shape-Based 3-D Scaffold Hopping Method and Its Application to a Bacterial Protein−Protein Interaction. J Med Chem 2005, 48, 1489-1495. (47) Sheridan, R. P.; Kearsley, S. K. Why Do We Need So Many Chemical Similarity S...

    2005 doi arXiv

  7. [21]

    P.; Hoeschele, J

    (140) Johnson, N. P.; Hoeschele, J. D.; Rahn, R. O.; O'Neill, J. P.; Hsie, A. W. Mutagenicity, Cytotoxicity, and DNA Binding of Platinum(Ii)-Chloroammines in Chinese Hamster Ovary Cells. Cancer Res 1980, 40, 1463-1468. (141) Rosenberg, B. Some Biological Effects of Platinum Co...

    1980

  8. [22]

    K.; Casewit, C

    (79) Rappé, A. K.; Casewit, C. J.; Colwell, K. S.; III, W. A. G.; Skiff, W. M. Uff, a Full Periodic Table Force Field for Molecular Mechanics and Molecular Dynamics Simulations. J Am Chem Soc 1992, 114, 10024-10035. (80) Mardirossian, N.; Head-Gordon, M. Wb97m-V: A Combinatori...

    1992

  9. [24]

    J.; Zunker, M.; Wolf, J

    (159) Gisdon, F. J.; Zunker, M.; Wolf, J. N.; Prufer, K.; Ackermann, J.; Welsch, C.; Koch, I. Graph-Theoretical Prediction of Biological Modules in Quaternary Structures of Large Protein Complexes. Bioinform 2024,

    2024

  10. [27]

    (96) Reichardt, J.; Bornholdt, S

    (95) RDKit; 2026.3.1 ed., 2026, DOI:10.5281/zenodo.19250388 10.5281/zenodo.19250388. (96) Reichardt, J.; Bornholdt, S. Statistical Mechanics of Community Detection. Phys Rev E 2006,

    2026 doi

  11. [31]

    L.; Bouldin, D

    (102) Davies, D. L.; Bouldin, D. W. A Cluster Separation Measure. IEEE Trans Pattern Anal Mach Intell 1979, PAMI-1, 224-227. (103) Caliński, T.; Harabasz, J. A Dendrite Method for Cluster Analysis. Commun Stat - Theory Methods 1974, 3, 1-27. (104) Rousseeuw, P. J. Silhouettes:...

    1979 doi

  12. [32]

    L.; Bouldin, D

    (16) Davies, D. L.; Bouldin, D. W. A Cluster Separation Measure. IEEE Trans Pattern Anal Mach Intell 1979, PAMI-1, 224-227. (17) Caliński, T.; Harabasz, J. A Dendrite Method for Cluster Analysis. Commun Stat - Theory Methods 1974, 3, 1-27

    1979

  13. [45]

    S.; Iyer, S

    (5) Rosen, A. S.; Iyer, S. M.; Ray, D.; Yao, Z.; Aspuru-Guzik, A.; Gagliardi, L.; Notestein, J. M.; Snurr, R. Q. Machine Learning the Quantum-Chemical Properties of Metal–Organic Frameworks for Accelerated Materials Discovery. Matter 2021, 4, 1578-1597. (6) Rosen, A. S.; Fung,...

    2021

  14. [46]

    T.; Stöhr, M.; Ganscha, S.; Unterthiner, T.; Maennel, H.; Kashubin, S.; Ahlin, D.; Gastegger, M.; Sandonas, L

    (7) Unke, O. T.; Stöhr, M.; Ganscha, S.; Unterthiner, T.; Maennel, H.; Kashubin, S.; Ahlin, D.; Gastegger, M.; Sandonas, L. M.; Berryman, J. T.; Tkatchenko, A.; Müller, K.-R. Page S16 Biomolecular Dynamics with Machine-Learned Quantum-Mechanical Force Fields Trained on Diverse...

    2024

  15. [47]

    (9) Terrones, G.; Michel, R

    (8) RDKit; 2026.3.1 ed., 2026, DOI:10.5281/zenodo.19250388 10.5281/zenodo.19250388. (9) Terrones, G.; Michel, R. S.; Toney, J.; Ball, A.; Wang, Y.; Garrison, A.; Nandy, A.; Meyer, R.; Edholm, F.; Oh, C.; Pujet, S.; Chu, D.; Muhammetgulyyev, D.; Kulik, H. Molsimplify 2.0: Impro...

    2026 doi arXiv

  16. [63]

    C.; Arriaga, E

    (158) Li, Y.; Nguyen, J.; Anastasiu, D. C.; Arriaga, E. A. Costal: An Accurate and Scalable Graph-Based Clustering Algorithm for High-Dimensional Single-Cell Data Analysis. Brief Bioinform 2023,

    2023

  17. [74]

    A.; Newman, M

    (97) Leicht, E. A.; Newman, M. E. J. Community Structure in Directed Networks. Phys Rev Lett 2008,

    2008

  18. [87]

    S.; Riley, P

    (27) Gilmer, J.; Schoenholz, S. S.; Riley, P. F.; Vinyals, O.; Dahl, G. E. Neural Message Passing for Quantum Chemistry. Proceedings of the 34th International Conference on Machine Learning 2017, 70, 1263-1272. (28) Satorras, V. c. G.; Hoogeboom, E.; Welling, M. In Proceedings...

    2017

  19. [98]

    P.; Kondor, R.; Csányi, G

    (26) Bartók, A. P.; Kondor, R.; Csányi, G. On Representing Chemical Environments. Phys Rev B 2013,

    2013

  20. [100]

    A.; Waltman, L.; van Eck, N

    (98) Traag, V. A.; Waltman, L.; van Eck, N. J. From Louvain to Leiden: Guaranteeing Well-Connected Communities. Sci Rep 2019,

    2019

  21. [108]

    Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces

    38 (25) Behler, J.; Parrinello, M. Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces. Phys Rev Lett 2007,

    2007

  22. [139]

    (29) Preuer, K.; Klambauer, G.; Rippmann, F.; Hochreiter, S.; Unterthiner, T. In Explainable Ai: Interpreting, Explaining and Visualizing Deep Learning; Samek, Wojciech;Montavon, Grégoire;Vedaldi, Andrea;Hansen, Lars Kai;Müller, Klaus-Robert, Eds.; Springer, Cham, 2019, DOI:10...

    2019 doi

  23. [539]

    (147) Yang, L.; Song, G.; Jernigan, R. L. Protein Elastic Network Models and the Ranges of Cooperativity. Proc Natl Acad Sci 2009, 106, 12347-12352. (148) Perez, A.; Yang, Z.; Bahar, I.; Dill, K. A.; MacCallum, J. L. Flexe: Using Elastic Network Models to Compare Models of Pro...

    2009

  24. [779]

    S.; Iyer, S

    (89) Rosen, A. S.; Iyer, S. M.; Ray, D.; Yao, Z.; Aspuru-Guzik, A.; Gagliardi, L.; Notestein, J. M.; Snurr, R. Q. Machine Learning the Quantum-Chemical Properties of Metal–Organic Frameworks for Accelerated Materials Discovery. Matter 2021, 4, 1578-1597. (90) Rosen, A. S.; Fun...

    2021

  25. [892]

    Geometry-Enhanced Molecular Representation Learning for Property Prediction

    (45) Fang, X.; Liu, L.; Lei, J.; He, D.; Zhang, S.; Zhou, J.; Wang, F.; Wu, H.; Wang, H. Geometry-Enhanced Molecular Representation Learning for Property Prediction. Nat Mach Intell 2022,

    2022

  26. [1695]

    F.; Zanini, F

    43 (117) Antonov, M.; Csárdi, G.; Horvát, S.; Müller, K.; Nepusz, T.; Noom, D.; Salmon, M.; Traag, V.; Welles, B. F.; Zanini, F. Igraph Enables Fast and Robust Network Analysis across Programming Languages. arXiv 2023, DOI:10.48550/arXiv.2311.10260 10.48550/arXiv.2311.10260. (...

    2023 doi arXiv

  27. [1967]

    (119) Lloyd, S. P. Least Squares Quantization in Pcm. IEEE Transactions on Information Theory 1982, 28, 129-137. (120) Sculley, D. WWW '10: Proceedings of the 19th international conference on World wide web, Raleigh, North Carolina, USA, 2010; p 1177-1178. (121) Pearson, K. On...

    1982

  28. [1990]

    Molecular Similarity in Medicinal Chemistry

    (39) Maggiora, G.; Vogt, M.; Stumpfe, D.; Bajorath, J. Molecular Similarity in Medicinal Chemistry. J Med Chem 2013, 57, 3186-3204. (40) López-Pérez, K.; Avellaneda-Tamayo, J. F.; Chen, L.; López-López, E.; Juárez-Mercado, K. E.; Medina-Franco, J. L.; Miranda-Quintana, R. A. M...

    2013

  29. [1991]

    The Principles Behind Equivariant Neural Networks for Physics and Chemistry

    (70) Kondor, R. The Principles Behind Equivariant Neural Networks for Physics and Chemistry. Proc Natl Acad Sci 2025,

    2025

  30. [1992]

    Chemical Graph Theory: A Combination of Chemistry and Maths

    (68) Wolf, R. Chemical Graph Theory: A Combination of Chemistry and Maths. Arch Chem Res 2022,

    2022

  31. [1997]

    Algebraic Connectivity of Graphs

    (72) Fiedler, M. Algebraic Connectivity of Graphs. Czechoslovak Mathematical Journal 1973, 23, 298–305. (73) Fiedler, M. Laplacian of Graphs and Algebraic Connectivity. Banach Center Publications 1987, 25, 57-70. (74) Cvetković, D.; Rowlinson, P.; Simić, S. An Introduction to ...

    1973

  32. [2003]

    ACM SIGR conference on Research and Development in information retrieval, 2004; p 96-103

    (161) He, X.; Cai, D.; Liu, H.; Ma, W.-Y. ACM SIGR conference on Research and Development in information retrieval, 2004; p 96-103. (162) Xie, D.; Zhang, X.; Gao, Q.; Han, J.; Xiao, S.; Gao, X. Multiview Clustering by Joint Latent Representation and Similarity Learning. IEEE T...

    2004

  33. [2004]

    The General Theory of Molecular Forces

    (130) London, F. The General Theory of Molecular Forces. Trans. Farad. Soc. 1937, 33, 8-26. (131) Miyato, T.; Kataoka, T.; Koyama, M.; Yoshida, Y. In 6th International Conference on Learning Representations Vancouver, Canada, 2018, DOI:10.48550/arXiv.1802.05957 10.48550/arXiv....

    1937 doi arXiv

  34. [2007]

    Gly-to-Ala mutant

    For Table of Contents Use Only 46 Page S1 Supporting Information for Physics-Based Molecular Fingerprints from Spectral Graph Theory Provide Efficient Geometry-Aware Measures of Chemical Similarity Jacob W. Toney1,2, Ayleen Y. Farnood1,2, Samir Darouich1,3,4, and Heather J. Ku...

    2026

  35. [2008]

    S.; Smith, K

    (76) Pearlman, R. S.; Smith, K. M. Metric Validation and the Receptor-Relevant Subspace Concept. J Chem Inf Model 1999, 39, 28-35. (77) Avogadro; 1.2.0 ed. (78) Hanwell, M. D.; Curtis, D. E.; Lonie, D. C.; Vandermeersch, T.; Zurek, E.; Hutchison, G. R. Avogadro: An Advanced Se...

    1999

  36. [2013]

    G.; Minyaev, R

    (136) Starikov, A. G.; Minyaev, R. M.; Minkin, V. I. Theoretical Modeling of the Square-Planar to Tetrahedral Isomerization of Bis-Chelate Nickel(Ii) Complexes. Chem Phys Lett 2008, 459, 27-32. (137) Lohbeck, K.; Haferkorn, H.; Fuhrmann, W.; Fedtke, N. In Ullmann's Encyclopedi...

    2008 doi

  37. [2015]

    (127) Schwerdtfeger, P.; Nagle, J. K. 2018 Table of Static Dipole Polarizabilities of the Neutral Elements in the Periodic Table. Mol Phys 2019, 117, 1200-1225. (128) Alvarez, S. A Cartography of the Van Der Waals Territories. Dalton Trans 2013, 42, 8617-36. (129) Cramer, C. J...

    2018

  38. [2022]

    Can One Hear the Shape of a Molecule (from Its Coulomb Matrix Eigenvalues)? J Chem Inf Model 2020, 60, 3804-3811

    (133) Schrier, J. Can One Hear the Shape of a Molecule (from Its Coulomb Matrix Eigenvalues)? J Chem Inf Model 2020, 60, 3804-3811. (134) Bruice, P. Y. Organic Chemistry; 8th ed.; Pearson: Hoboken, New Jersey, USA,

    2020

  39. [2775]

    S.; Pouyamehr, F.; Koohi, S

    (150) Akbari Rokn Abadi, S.; Abdosalehi, A. S.; Pouyamehr, F.; Koohi, S. An Accurate Alignment-Free Protein Sequence Comparator Based on Physicochemical Properties of Amino Acids. Sci Rep 2022, 12, 11158. (151) Röhling, S.; Linne, A.; Schellhorn, J.; Hosseini, M.; Dencker, T.;...

    2022

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.