Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Persistent Homology: A Pedagogical Introduction with Biological Applications

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Persistent homology, applied to a 3-1 supercoiled DNA structure, reduces its shape to Betti numbers β0=1, β1=3, β2=0 — one connected component, three significant loops, and no voids.

desk verdict A clear, correct pedagogical survey of persistent homology whose biological application is compromised by a protein/DNA mix-up and a contradictory interpretation of H1 significance; fixable, and worth fixing. read the letter →

arxiv 2505.06583 v1 pith:G3KYAFV6 submitted 2025-05-10 math.AT

Aurelie Jodelle Kemme , Collins Amburo Agyingi This is my paper · ORCID
classification math.AT MSC 55N3155-0192-08
keywords persistenthomologytopologicaldataanalysisBettinumberssupercoiledDNAVietoris–Ripsfiltrationpersistencebarcodepedagogicalintroductionproteinstructure
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 sets out to make persistent homology accessible to readers with no prior algebraic topology, and to demonstrate it on a real biological object. It builds the pipeline from first principles — topological spaces, simplicial complexes, chain groups, homology, Betti numbers — and then applies it to a 3-1 supercoiled DNA structure. The worked example reports Betti numbers β0=1, β1=3, β2=0, which the authors read as one connected component, three significant loops or cycles, and no enclosed voids. The broader claim is that this multi-scale topological summary captures structure that ordinary machine learning on coordinates would miss, and that the method is learnable by nonspecialists.

What carries the argument

The carrying object is the Vietoris–Rips filtration built from pairwise Euclidean distances among the coordinate points. As the scale radius grows, simplices are added and topological features are born and die; homology in each dimension counts cycles modulo boundaries. The output is a persistence diagram and a barcode, whose long bars designate the meaningful features, and the final numeric summary is the set of Betti numbers. The structure theorem for persistence modules justifies encoding persistent homology as a multiset of intervals, and the stability theorem supports treating long-lived features as reliable.

What would settle it

Fetch the structure used in the example and inspect its database entry: if the sequence is a protein rather than the named 3-1m-supercoiled DNA, the point cloud is not a faithful DNA representation. Recomputing the Vietoris–Rips persistence diagram on a correctly labelled DNA double-helix model would settle whether β1=3 is a genuine property of supercoiling or an artefact of the coordinate source.

Watch

Extended reading notes

Core claim

The paper's central claim is that a 3-1m-supercoiled DNA molecule, represented as a point cloud of alpha-carbon coordinates and filtered with a Vietoris–Rips construction, carries the persistent homology signature β0=1, β1=3, β2=0. The intended reading is that the molecule is a single connected piece with three long-lived one-dimensional cycles — twists or knot-like loops that may be biologically meaningful — and no cavities. Around this example the paper also claims that the full persistent-homology pipeline, from pairwise distances to barcodes to Betti numbers, is usable by people with no prior topological training.

Load-bearing premise

The whole application rests on treating the downloaded 3D coordinates as a valid protein-style model of the supercoiled DNA molecule; if those coordinates are mislabelled or the alpha-carbon convention does not apply to DNA, the Betti numbers describe a different object.

Editorial extensions

If this is right

  • Persistent homology reduces a complex three-dimensional biomolecular geometry to a small set of interpretable numbers that can serve as features in clustering or classification pipelines.
  • Because long-lived topological features are stable under small perturbations of the input coordinates, the three reported loops are more trustworthy than short-lived fluctuations.
  • The same step-by-step recipe — point cloud, filtration, barcode, Betti numbers — transfers directly to other biological structures such as proteins, RNA, or molecular dynamics ensembles.
  • For nonspecialists, the tutorial format lowers the barrier to applying topological data analysis without first taking a course in algebraic topology.
  • The computed signature (one component, three loops, zero voids) gives a concrete topological fingerprint of the supercoiled state that could be compared across molecular conformations.

Reading between the lines

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

  • My inference: if the coordinate set is valid but the molecule is actually a protein rather than the named DNA, then the reported 'DNA' signature is in fact a protein-folding signature; re-labelling the object preserves the computation but changes the biological interpretation.
  • My inference: the three one-dimensional cycles likely correspond to superhelical turns or plectonemic crossings; comparing persistence diagrams of relaxed versus supercoiled forms of the same sequence would test whether β1 tracks supercoiling density.
  • My inference: the paper's own tutorial pipeline suggests a natural extension it does not run — vectorising the persistence diagram into a persistence landscape or persistence image and feeding it to a classifier to distinguish topoisomers automatically.
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

3 major / 4 minor

Summary. The paper presents a pedagogical introduction to persistent homology, moving from topological spaces, homeomorphisms, homotopy, and metric spaces through simplicial complexes, homology, and persistence diagrams/barcodes, and then applies the pipeline to a structure labeled '3-1m-supercoiled DNA'. The application extracts xyz coordinates of carbon-alpha atoms from AlphaFold, computes a Vietoris–Rips filtration with GUDHI/Ripser, and reports Betti numbers β0=1, β1=3, β2=0, interpreting β1=3 as three significant loops or cycles.

Significance. The pedagogical sections (Sections 2–4) contain mostly standard, correctly stated definitions and will likely be useful to readers without an algebraic-topology background. The paper also provides concrete implementation pointers and a worked computational example, which increase its value as an introductory tutorial. However, the application section contains a load-bearing inconsistency about the biological object being analyzed, and the interpretation of the H1 features is internally contradictory. Because the stated application is the paper's main demonstration, these issues need to be resolved before the manuscript can be accepted.

major comments (3)
  1. [§5, Step 1 and §5.1] The application's input object is described inconsistently. Section 5 opens by stating that the method is applied to a protein, but §5.1 identifies the same structure as '3-1m-supercoiled DNA molecules' and says its 3D structure was downloaded from the AlphaFold database using carbon-alpha (Cα) atom coordinates. AlphaFold contains protein structures, and Cα atoms are protein backbone atoms; DNA has no amino acids and no Cα atoms. As written, the reported Betti numbers in Table 3 cannot be claimed to describe supercoiled DNA. The authors must either correct the biological object and data source or explicitly re-frame the example as a protein-structure analysis.
  2. [§5.1, Figure 15 and Table 3] The interpretation of the H1 features is internally contradictory. The text accompanying Figure 15 says that H1 features have 'relatively short lifespans, suggesting minor or less significant loops,' while the discussion after Table 3 says that β1=3 reveals 'the presence of three significant loops or cycles.' Since the significance of loops is the central biological conclusion of the application, these two statements must be reconciled with a concrete persistence criterion.
  3. [§5.1] The computation is not reproducible from the manuscript as written. No filtration parameters, persistence pairs, or code output are given—only the final Betti numbers. The claim that the three H1 features are 'significant' therefore rests on an unspecified threshold. Please include the persistence diagram coordinates or a stated persistence threshold that justifies labeling β1=3 as three significant cycles.
minor comments (4)
  1. [§5] The terms 'protein' and 'DNA' are used interchangeably; the authors should choose a consistent terminology for the analyzed object.
  2. [§5.1] The heading 'Vietoris-Risp filtration' should read 'Vietoris–Rips filtration'.
  3. [Figure 12] The caption contains the typo 'residus'; it should be 'residues'.
  4. [Definition 11] The boundary maps are denoted δ_k earlier, but Definition 11 uses d_k; the notation should be unified.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the Betti numbers are computed from input coordinates via a standard pipeline, with no fitted parameters or self-referential derivation.

full rationale

The paper is a pedagogical exposition of persistent homology followed by a worked computation on a point cloud derived from atomic coordinates. The claimed results—β0=1, β1=3, β2=0—are outputs of a Vietoris–Rips filtration computed with Ripser/GUDHI from the (x,y,z) coordinates, not predictions obtained by fitting parameters to reproduce a target. No step in the exposition defines a concept in terms of the result it is supposed to derive, and no load-bearing premise is justified solely by a self-citation. The AlphaFold/Cα procedure is described as data acquisition, not as a derivation of the topological conclusion, and the biological interpretation of the Betti numbers is post-hoc commentary rather than an input to the computation. The paper's internal inconsistency about whether the molecule is DNA or protein, and the tension between 'short lifespans' in the barcode text and 'significant loops' in the Betti-number interpretation, are correctness or clarity issues, not circularity. The derivation chain is therefore self-contained with respect to circularity concerns.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new mathematical entities or fitted parameters. It relies on standard theory and on the biological validity of the AlphaFold-derived point cloud, which is not adequately justified. The assumption about software correctness is also load-bearing because the central application is a computational result.

assumptions (3)
  • domain assumption The topological features computed by the Ripser/GUDHI implementation correctly reflect the homology of the Vietoris-Rips filtration.
    The paper relies on these software packages to produce the Betti numbers in Table 3, but it does not validate the implementations or the specific parameter settings used.
  • domain assumption The AlphaFold-predicted structure of the '3-1m-supercoiled' molecule is a valid biological representation, and alpha-carbon coordinates are appropriate for this analysis.
    This assumption is introduced in Section 5, Step 1, where the paper uses Cα atoms from AlphaFold for a molecule it calls DNA, without explaining the validity of this representation.
  • standard math The structure theorem for persistence modules and the stability theorem hold for finite filtered simplicial complexes.
    These are standard results cited from Chazal et al. (2016) and Oudot (2015), and the paper invokes them to justify interpreting barcodes as stable summaries.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Persistent Homology: A Pedagogical Introduction with Biological Applications." pith.science (2026). https://pith.science/paper/G3KYAFV6

@misc{pith2026250506583,
  author       = {Pith},
  title        = {Pith review of: Persistent Homology: A Pedagogical Introduction with Biological Applications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G3KYAFV6}},
  note         = {Machine review of arXiv:2505.06583}
}
read the original abstract

Persistent Homology (PH) is a fundamental tool in computational topology, designed to uncover the intrinsic geometric and topological features of data across multiple scales. Originating within the broader framework of Topological Data Analysis (TDA), PH has found diverse applications ranging from protein structure and knot analysis to financial domains such as Bitcoin behaviour and stock market dynamics. Despite its growing relevance, there remains a lack of accessible resources that bridge the gap between theoretical foundations and practical implementation for beginners. This paper offers a clear and comprehensive introduction to persistent homology, guiding readers from core concepts to real-world applications. Specifically, we illustrate the methodology through the analysis of a 3-1 supercoiled DNA structure. The paper is tailored for readers without prior exposure to algebraic topology, aiming to demystify persistent homology and foster its broader adoption in data analysis tasks.

Figures

Figures reproduced from arXiv: 2505.06583 by the authors.

Figure 1
Figure 1. Some topologically equivalent objects. source: Reese Cabrera [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. A coffee mug and a donut are homotopically equivalent. source: Marlowe Hood, GMA News [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Visualization of three different types of point clouds. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: Some metrics commonly used in Machine Learning. source: Maarten Grootendorst [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Example of simplicial complexes The [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: A Simplex with some of its faces The [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 8
Figure 8. Figure 8: Example of boundary of an object Chain Complexes. At a high level, a chain complex is a mathematical structure that helps us organize and analyze the different building blocks (simplices) of a simplicial complex across various dimensions. It provides a systematic way t…
Figure 9
Figure 9. Figure 9: Visual Representation of Common Shapes and Their Topological Features [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: x,y,z coordinates of 3-1m-Supercoiled The figure 10 is a table with 3 columns and 111 rows. Where the columns are respectively the x, y, and z coordinates, and the number of rows represents the number of points used to represent the 3D structure of the 3-1m-Supercoile…
Figure 11
Figure 11. Figure 11: Point cloud of the 3-1m-Supercoiled. From the figure 11, we can see each data point with its corresponding x, y, z coordinates, leading us to the notion of neighbourhood. We want to measure how data points are close or far from each other, and for that, we will start …
Figure 12
Figure 12. Figure 12: Pairwise distances illustrating proximity between residus [PITH_FULL_IMAGE:figures/full_fig_p022_12.png]
Figure 13
Figure 13. Figure 13: Initial Stages of Vietoris–Rips Filtration: Early Edge Formation and Emerging Loops. [PITH_FULL_IMAGE:figures/full_fig_p022_13.png]
Figure 14
Figure 14. Figure 14: Vietoris–Rips Filtration Revealing the Birth and Death of 1-Dimensional Homological Features. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_14.png]
Figure 15
Figure 15. Figure 15: Persistence Diagram (left) and Persistence Barcode (right) showing topological features across scales [PITH_FULL_IMAGE:figures/full_fig_p024_15.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Neural-Accelerated Bayesian Calibration of Chiral Mean-Field Models to Nuclear Saturation and Vacuum Properties

    nucl-th 2026-07 conditional novelty 6.0 of 10

    A neural-accelerated Bayesian calibration of the chiral mean-field model shows that nuclear vacuum and saturation data constrain combinations of couplings while leaving individual parameters and neutron-star predictio...

Reference graph

Works this paper leans on

30 extracted references · 20 canonical work pages · cited by 1 Pith paper

  1. [1]

    Persistenceimages: Astablevectorrepresentation of persistent homology.Journal of Machine Learning Research, 18(8):1–35, 2017

    HenryAdamsandTamalK.etal.Emerson. Persistenceimages: Astablevectorrepresentation of persistent homology.Journal of Machine Learning Research, 18(8):1–35, 2017

    work page 2017

  2. [2]

    Tight basis cycle representatives for persistent homology of large biological data sets.PLOS Computational Biology, 19(5):e1010341, 2023

    Manu Aggarwal and Vipul Periwal. Tight basis cycle representatives for persistent homology of large biological data sets.PLOS Computational Biology, 19(5):e1010341, 2023

    work page 2023

  3. [3]

    Statistical topological data analysis using persistence landscapes.Journal of Machine Learning Research, 16(1):77–102, 2015

    Peter Bubenik. Statistical topological data analysis using persistence landscapes.Journal of Machine Learning Research, 16(1):77–102, 2015

    2015

  4. [4]

    Topology and data.Bulletin of the American Mathematical Society, 46(2): 255–308, 2009

    Gunnar Carlsson. Topology and data.Bulletin of the American Mathematical Society, 46(2): 255–308, 2009

    2009

  5. [5]

    Topological data analysis and machine learning theory

    Gunnar Carlsson, Rick Jardine, Dmitry Feichtner-Kozlov, Dmitriy Morozov, Frédéric Chazal, Vin de Silva, Brittany Fasy, Jesse Johnson, Matt Kahle, Gilad Lerman, et al. Topological data analysis and machine learning theory. InProc. BIRS Workshop, pages 1–11, 2012

    work page 2012

  6. [6]

    Springer, 2016

    Frédéric Chazal, Vin De Silva, Marc Glisse, and Steve Oudot.The structure and stability of persistence modules, volume 10. Springer, 2016

    work page 2016

  7. [7]

    Stability of persistence diagrams.Discrete & Computational Geometry, 37(1):103–120, 2007

    David Cohen-Steiner, Herbert Edelsbrunner, and John Harer. Stability of persistence diagrams.Discrete & Computational Geometry, 37(1):103–120, 2007. 25

    work page 2007

  8. [8]

    American Mathematical Society, 2010

    Herbert Edelsbrunner and John Harer.Computational Topology: An Introduction. American Mathematical Society, 2010

    2010

Show all 30 references
  1. [9]

    Topological persistence and simplification.Discrete & Computational Geometry, 28(4):511–533, 2002

    Herbert Edelsbrunner, David Letscher, and Afra Zomorodian. Topological persistence and simplification.Discrete & Computational Geometry, 28(4):511–533, 2002

    2002

  2. [10]

    A density-based algorithm for discovering clusters in large spatial databases with noise

    Martin Ester, Hans-Peter Kriegel, Jörg Sander, and Xiaowei Xu. A density-based algorithm for discovering clusters in large spatial databases with noise. InProceedings of the Second International Conference on Knowledge Discovery and Data Mining (KDD), pages 226–231, 1996

    1996

  3. [11]

    No Starch Press, 2023

    Colleen M Farrelly and Yaé Ulrich Gaba.The Shape of Data: Geometry-Based Machine Learning and Data Analysis in R. No Starch Press, 2023

    2023

  4. [12]

    Barcodes: The persistent topology of data.Bulletin of the American Mathematical Society, 45(1):61–75, 2008

    Robert Ghrist. Barcodes: The persistent topology of data.Bulletin of the American Mathematical Society, 45(1):61–75, 2008

    2008

  5. [13]

    Topological data analysis of financial time series: Landscapes of crashes.Physica A: Statistical Mechanics and its Applications, 491:820–834, 2018

    Marian Gidea and Yuri Katz. Topological data analysis of financial time series: Landscapes of crashes.Physica A: Statistical Mechanics and its Applications, 491:820–834, 2018

    2018

  6. [14]

    Chad Giusti, Robert Ghrist, and Danielle S. Bassett. Two’s company, three (or more) is a simplex: Algebraic-topological tools for understanding higher-order structure in neural data. Journal of Computational Neuroscience, 41(1):1–14, 2016

    2016

  7. [15]

    MIT Press, 2016

    Ian Goodfellow, Yoshua Bengio, and Aaron Courville.Deep Learning. MIT Press, 2016

    2016

  8. [16]

    Unsupervised machine learning.Bioconductor case studies, pages 137–157, 2008

    FlorianHahne, WolfgangHuber, RobertGentleman, SethFalcon, RGentleman, andVJCarey. Unsupervised machine learning.Bioconductor case studies, pages 137–157, 2008

    2008

  9. [17]

    Reinforcement learning: A survey.Journal of artificial intelligence research, 4:237–285, 1996

    Leslie Pack Kaelbling, Michael L Littman, and Andrew W Moore. Reinforcement learning: A survey.Journal of artificial intelligence research, 4:237–285, 1996

    1996

  10. [18]

    Peter M. Kasson. Topological data analysis of protein folding dynamics.Current Opinion in Structural Biology, 60:138–144, 2020

    2020

  11. [19]

    Supervised machine learning approaches: A survey.ICTACT Journal on Soft Computing, 5(3), 2015

    Iqbal Muhammad and Zhu Yan. Supervised machine learning approaches: A survey.ICTACT Journal on Soft Computing, 5(3), 2015

    2015

  12. [20]

    An unsupervised machine learning algorithms: Comprehensive review.International Journal of Computing and Digital Systems, 2023

    Samreen Naeem, Aqib Ali, Sania Anam, and Muhammad Munawar Ahmed. An unsupervised machine learning algorithms: Comprehensive review.International Journal of Computing and Digital Systems, 2023

    2023

  13. [21]

    Computational algebraic topology lecture notes.URL: https://people

    Vidit Nanda. Computational algebraic topology lecture notes.URL: https://people. maths. ox. ac. uk/nanda/cat/TDANotes. pdf, 2021

    2021

  14. [22]

    An overview of the supervised machine learning methods.Horizons

    Vladimir Nasteski. An overview of the supervised machine learning methods.Horizons. b, 4 (51-62):56, 2017

    2017

  15. [23]

    A roadmap for the computation of persistent homology.EPJ Data Science, 6(1):17, 2017

    Nina Otter, Mason A Porter, Ulrike Tillmann, Peter Grindrod, and Heather A Harrington. A roadmap for the computation of persistent homology.EPJ Data Science, 6(1):17, 2017

    2017

  16. [24]

    Oudot.Persistence Theory: From Quiver Representations to Data Analysis, volume 209 ofMathematical Surveys and Monographs

    Steve Y. Oudot.Persistence Theory: From Quiver Representations to Data Analysis, volume 209 ofMathematical Surveys and Monographs. American Mathematical Society, 2015. 26

    2015

  17. [25]

    Topological data analysis in graph neural networks: surveys and perspectives.IEEE Transactions on Neural Networks and Learning Systems, 2025

    Phu Pham, Quang-Thinh Bui, Ngoc Thanh Nguyen, Robert Kozma, Philip S Yu, and Bay Vo. Topological data analysis in graph neural networks: surveys and perspectives.IEEE Transactions on Neural Networks and Learning Systems, 2025

    2025

  18. [26]

    Persistent-homology-based machine learning and its applications – a survey, 2018

    Chi Seng Pun, Kelin Xia, and Si Xian Lee. Persistent-homology-based machine learning and its applications – a survey, 2018. URLhttps://arxiv.org/abs/1811.00252

    2018 arXiv

  19. [27]

    Cycle registration in persistent homology with applications in topological bootstrap.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(5):5579–5593, 2022

    Yohai Reani and Omer Bobrowski. Cycle registration in persistent homology with applications in topological bootstrap.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(5):5579–5593, 2022

    2022

  20. [28]

    Explaining the power of topological data analysis in graph machine learning.arXiv preprint arXiv:2401.04250, 2024

    Funmilola Mary Taiwo, Umar Islambekov, and Cuneyt Gurcan Akcora. Explaining the power of topological data analysis in graph machine learning.arXiv preprint arXiv:2401.04250, 2024

    2024 arXiv

  21. [29]

    Alphafold protein structure database: massively expanding the structural coverage of protein-sequence space with high-accuracy models.Nucleic acids research, 50(D1):D439–D444, 2022

    Mihaly Varadi, Stephen Anyango, Mandar Deshpande, Sreenath Nair, Cindy Natassia, Galabina Yordanova, David Yuan, Oana Stroe, Gemma Wood, Agata Laydon, et al. Alphafold protein structure database: massively expanding the structural coverage of protein-sequence space with high-a...

    2022

  22. [30]

    Computing persistent homology.Discrete & Computational Geometry, 33(2):249–274, 2005

    Afra Zomorodian and Gunnar Carlsson. Computing persistent homology.Discrete & Computational Geometry, 33(2):249–274, 2005. 27

    2005

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.