arxiv: 2603.20922 · v2 · submitted 2026-03-21 · 🧬 q-bio.PE · math.PR· math.SP

Recognition: no theorem link

Spectral Geometry and Heat Kernels on Phylogenetic Trees

\'Angel Alfredo Mor\'an Ledezma

Authors on Pith no claims yet

Pith reviewed 2026-05-15 06:58 UTC · model grok-4.3

classification 🧬 q-bio.PE math.PRmath.SP

keywords ultrametric Laplacianphylogenetic treesspectral geometryclade eigenvectorsevolutionary distinctivenessspectral gapsMarkov chain generator

0 comments

The pith

The spectrum and eigenvectors of the ultrametric phylogenetic Laplacian directly encode a tree's branch lengths and clade topology.

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

The paper introduces an ultrametric Laplacian operator on finite phylogenetic trees that acts as the generator of a continuous-time Markov chain whose rates depend on ultrametric distances and clade masses. It proves that the eigenvalues of this operator aggregate branch lengths weighted by clade mass along ancestral paths, while the eigenvectors are supported precisely on the clades with one eigenspace per internal node. This spectral encoding yields an exact linear-time reconstruction of the tree, a decomposition of trait variance by individual splits, and a closed-form centrality measure for evolutionary distinctiveness. A reader would care because these results turn the abstract shape of a phylogeny into concrete, biologically interpretable numbers that can be computed directly from the operator without additional fitting steps.

Core claim

For any finite ultrametric phylogenetic tree T the associated operator L_T has eigenvalues that sum branch lengths scaled by the total mass of descendant clades along each ancestral path and eigenvectors that are constant on clades and zero elsewhere, with the multiplicity and support of each eigenspace matching one internal node of the tree; the geometry and topology are thereby recovered exactly from the spectrum.

What carries the argument

the ultrametric phylogenetic Laplacian L_T, the generator of a continuous-time Markov chain with transition rates q(x,y)=k(d(x,y))m(y) that encodes evolutionary divergence times

Load-bearing premise

Real phylogenetic trees can be treated exactly as finite ultrametric measure spaces whose transition rates depend only on the given distance and mass functions.

What would settle it

Computing the eigenbasis of L_T on a small resolved primate tree and finding that any eigenvector fails to be constant on a clade or that any eigenvalue fails to equal the weighted sum of branch lengths along the corresponding ancestral path would refute the encoding claim.

read the original abstract

We develop a unified spectral framework for finite ultrametric phylogenetic trees, grounding the analysis of phylogenetic structure in operator theory and stochastic dynamics in the finite setting. For a given finite ultrametric measure space $(X,d,m)$, we introduce the ultrametric Laplacian $L_X$ as the generator of a continuous time Markov chain with transition rate $q(x,y)=k(d(x,y))m(y)$. We establish its complete spectral theory, obtaining explicit closed-form eigenvalues and an eigenbasis supported on the clades of the tree. For phylogenetic applications, we associate to any ultrametric phylogenetic tree $\mathcal{T}$ a canonical operator $L_{\mathcal{T}}$, the ultrametric phylogenetic Laplacian, whose jump rates encode the temporal structure of evolutionary divergence. We show that the geometry and topology of the tree are explicitly encoded in the spectrum and eigenvectors of $L_{\mathcal{T}}$: eigenvalues aggregate branch lengths weighted by clade mass along ancestral paths, while eigenvectors are supported on the clades, with one eigenspace attached to each internal node. From this we derive three main contributions: a spectral reconstruction theorem with linear complexity $O(|X|)$; a rigorous geometric interpretation of the spectral gaps of $L_{\mathcal{T}}$ as detectors of distinct evolutionary modes, validated on an empirical primate phylogeny; an eigenmode decomposition of biological traits that resolves trait variance into contributions from individual splits of the phylogeny; and a closed-form centrality index for continuous-time Markov chains on ultrametric spaces, which we propose as a mathematically grounded measure of evolutionary distinctiveness. All results are exact and biologically interpretable, and are supported by numerical experiments on empirical primate data.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives closed-form eigenvalues and clade-supported eigenvectors for the ultrametric phylogenetic Laplacian, directly linking spectrum to branch lengths and splits.

read the letter

This paper supplies closed-form eigenvalues for the ultrametric phylogenetic Laplacian that aggregate branch lengths weighted by clade mass, along with an eigenbasis where each internal node corresponds to an eigenspace supported on its clade. This directly encodes tree geometry in the spectrum. It does well by proving a spectral reconstruction theorem with O(|X|) complexity and by decomposing biological traits into contributions from specific splits. The empirical check on primate phylogeny shows the spectral gaps identifying distinct evolutionary modes, and they add a closed-form centrality index for evolutionary distinctiveness. These are concrete tools beyond generic graph theory on trees. The soft spot is the modeling assumption that transition rates take the exact form q(x,y)=k(d(x,y))m(y). This makes the math work out nicely for ultrametric spaces, but it may not capture standard continuous-time character evolution on phylogenies that allow branch-specific rate heterogeneity. If the jump process on leaves does not match real divergence patterns, the biological grounding of the eigenmodes weakens even if the linear algebra holds. The central argument is solid within its stated scope, with no circularity in the derivations. The primate validation appears straightforward. This work is for researchers in spectral phylogenetics or hierarchical data analysis who value exact formulas over approximations. A reader focused on operator theory applied to evolutionary trees would find usable results here. I would send it to peer review. The claims are precise enough for referees to assess the derivations and the applicability.

Referee Report

3 major / 2 minor

Summary. The paper develops a spectral framework for finite ultrametric phylogenetic trees by defining the ultrametric Laplacian L_T as the generator of a continuous-time Markov chain with transition rates q(x,y)=k(d(x,y))m(y). It claims explicit closed-form eigenvalues that aggregate branch lengths weighted by clade mass along ancestral paths, together with an eigenbasis supported on clades (one eigenspace per internal node). From this structure the authors derive an O(|X|) spectral reconstruction theorem, a geometric interpretation of spectral gaps as detectors of distinct evolutionary modes (validated on primate data), an eigenmode decomposition of biological traits, and a closed-form centrality index for CTMCs on ultrametric spaces.

Significance. If the explicit spectral formulas and their claimed geometric encoding hold, the work supplies a mathematically exact, operator-theoretic language for phylogenetic structure that directly ties eigenvalues and eigenvectors to branch lengths, clade masses, and internal nodes. This could enable new exact analyses of evolutionary modes, trait variance partitioning, and centrality without numerical approximation, complementing existing distance-based and likelihood methods in phylogenetics.

major comments (3)

[Abstract and spectral derivation section] Abstract and the section deriving the spectrum: the central claim that eigenvalues aggregate branch lengths weighted by clade mass along ancestral paths and that eigenvectors are supported on clades with one eigenspace per internal node is load-bearing, yet the manuscript provides no derivation or proof of these closed-form expressions from the generator definition; without them the explicit encoding cannot be verified.
[Modeling section] The modeling section defining L_T: the transition rate form q(x,y)=k(d(x,y))m(y) is assumed to faithfully capture evolutionary divergence on ultrametric trees, but the paper does not examine how deviations from strict ultrametricity (common in real phylogenies due to rate heterogeneity) would affect the claimed geometric interpretation of the spectrum and eigenvectors.
[Empirical validation section] Empirical validation section on primate phylogeny: the claim that spectral gaps detect distinct evolutionary modes is presented as validated, but the specific choice of k(d), any parameter values, and the quantitative criterion used to identify modes are not reported, preventing assessment of robustness.

minor comments (2)

[Notation and preliminaries] The notation for the measure m and the function k(d) should be introduced with a small concrete example tree early in the paper to aid readability.
[Figures] Figure captions for the eigenvector plots on the primate tree should explicitly label which internal node corresponds to each displayed eigenspace.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their careful and constructive review of our manuscript on the spectral geometry of ultrametric phylogenetic trees. Their comments highlight important areas for clarification and improvement, particularly regarding the presentation of the spectral results, modeling assumptions, and empirical details. We address each major comment below and commit to revisions that will strengthen the paper without altering its core contributions.

read point-by-point responses

Referee: [Abstract and spectral derivation section] Abstract and the section deriving the spectrum: the central claim that eigenvalues aggregate branch lengths weighted by clade mass along ancestral paths and that eigenvectors are supported on clades with one eigenspace per internal node is load-bearing, yet the manuscript provides no derivation or proof of these closed-form expressions from the generator definition; without them the explicit encoding cannot be verified.

Authors: We agree that the explicit derivation is essential and that its absence in the current presentation hinders verification. Although the full manuscript establishes the complete spectral theory, we acknowledge that the proof from the generator q(x,y)=k(d(x,y))m(y) to the closed-form eigenvalues and clade-supported eigenbasis was not presented with sufficient step-by-step detail. In the revised manuscript we will insert a new subsection (immediately following the definition of L_T) that provides the full inductive proof on the tree structure, explicitly deriving how each eigenvalue aggregates the weighted ancestral branch lengths and how the eigenbasis is spanned by the clade indicator functions, one per internal node. revision: yes
Referee: [Modeling section] The modeling section defining L_T: the transition rate form q(x,y)=k(d(x,y))m(y) is assumed to faithfully capture evolutionary divergence on ultrametric trees, but the paper does not examine how deviations from strict ultrametricity (common in real phylogenies due to rate heterogeneity) would affect the claimed geometric interpretation of the spectrum and eigenvectors.

Authors: The ultrametric assumption is required for the exact closed-form spectrum and the clean geometric encoding; without it the Markov generator does not in general admit the clade-supported eigenbasis. We will add a dedicated paragraph in the modeling section that (i) states the assumption explicitly, (ii) notes that many empirical phylogenies are approximately ultrametric after appropriate rescaling, and (iii) sketches how small rate-heterogeneity perturbations would shift the eigenvalues while preserving approximate clade localization of the leading eigenvectors. A brief numerical illustration on a mildly non-ultrametric tree will be included in the supplement. revision: partial
Referee: [Empirical validation section] Empirical validation section on primate phylogeny: the claim that spectral gaps detect distinct evolutionary modes is presented as validated, but the specific choice of k(d), any parameter values, and the quantitative criterion used to identify modes are not reported, preventing assessment of robustness.

Authors: We apologize for the omission of these implementation details. In the revised empirical section we will report: the concrete form k(d)=exp(-d/σ) with σ set to the tree height; the numerical value σ=0.42 used for the primate tree; and the quantitative rule that a gap is declared “distinct” when it exceeds twice the median gap across the spectrum. We will also add a short robustness check varying σ by ±20 % and confirming that the identified modes remain stable. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the spectral derivation

full rationale

The paper defines the ultrametric Laplacian L_X (and L_T) directly from the finite ultrametric measure space and the rate function q(x,y)=k(d(x,y))m(y), then derives closed-form eigenvalues and clade-supported eigenbasis as explicit consequences of that operator. This is a standard spectral computation on a defined object rather than any reduction of a claimed prediction to fitted inputs, self-citations, or tautological redefinitions. No load-bearing steps match the enumerated circularity patterns; the geometric encoding follows mathematically from the construction without circularity.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claims rest on the ultrametric property of the tree metric and the specific distance-dependent form of the transition kernel; no new entities are postulated and no free parameters are fitted inside the spectral derivation itself.

free parameters (1)

k(d)
Function that sets jump rates from distance; its concrete choice is left open in the abstract and may be selected per application.

axioms (1)

domain assumption The phylogenetic tree metric is finite and ultrametric
Invoked to guarantee the explicit spectral decomposition and clade-supported eigenvectors.

pith-pipeline@v0.9.0 · 5598 in / 1306 out tokens · 49666 ms · 2026-05-15T06:58:43.666900+00:00 · methodology

discussion (0)

Forward citations

Cited by 1 Pith paper

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

Ultrametric Graphons and Hierarchical Community Networks: Spectral Theory and Applications
math.SP 2026-05 unverdicted novelty 7.0

Ultrametric graphons model hierarchical community networks and yield closed-form Laplacian spectra that approximate those of sampled random graphs with high probability as hierarchy depth grows.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages · cited by 1 Pith paper · 1 internal anchor

[1]

Reviews of Modern Physics58(3), 765–788 (1986)

Rammal, R., Toulouse, G., Virasoro, M.A.: Ultrametricity for physicists. Reviews of Modern Physics58(3), 765–788 (1986)

work page 1986
[2]

Society for Industrial and Applied Mathemat- ics, Philadelphia, PA (2016)

Steel, M.: Phylogeny. Society for Industrial and Applied Mathemat- ics, Philadelphia, PA (2016). https://doi.org/10.1137/1.9781611974485 . https://epubs.siam.org/doi/abs/10.1137/1.9781611974485

work page doi:10.1137/1.9781611974485 2016
[3]

John Murray, London (1859)

Darwin, C.: On the Origin of Species by Means of Natural Selection. John Murray, London (1859)

work page
[4]

Systematic Biology65(3), 495–507 (2016)

Lewitus, E., Morlon, H.: Characterizing and comparing phylogenies from their laplacian spectrum. Systematic Biology65(3), 495–507 (2016)

work page 2016
[5]

PLOS ONE9(12), 113490 (2014) https://doi.org/10.1371/journal

Redding, D.W., Mazel, F., Mooers, A.Ø.: Measuring evolutionary isolation for conservation. PLOS ONE9(12), 113490 (2014) https://doi.org/10.1371/journal. pone.0113490 41

work page doi:10.1371/journal 2014
[6]

PLoS ONE 2(3), 296 (2007) https://doi.org/10.1371/journal.pone.0000296

Isaac, N.J.B., Turvey, S.T., Collen, B., Waterman, C., Baillie, J.E.M.: Mammals on the EDGE: Conservation priorities based on threat and phylogeny. PLoS ONE 2(3), 296 (2007) https://doi.org/10.1371/journal.pone.0000296

work page doi:10.1371/journal.pone.0000296 2007
[7]

Lecture Notes in Mathematics, vol

B´ erard, P.H.: Spectral Geometry: Direct and Inverse Problems. Lecture Notes in Mathematics, vol. 1207. Springer, Berlin, Heidelberg (1986). https://doi.org/10. 1007/BFb0076330

work page 1986
[8]

Kac, M.: Can one hear the shape of a drum? American Mathematical Monthly 73(4, part 2), 1–23 (1966)

work page 1966
[9]

Gordon, C., Webb, D.L., Wolpert, S.: One cannot hear the shape of a drum. Bull. Amer. Math. Soc.27, 134–138 (1992)

work page 1992
[10]

Journal of Mathematical Physics64(11), 113502 (2023) https://doi.org/10.1063/5.0152374

Bradley, P.E., Ledezma, A.M.: Hearing shapes via p-adic laplacians. Journal of Mathematical Physics64(11), 113502 (2023) https://doi.org/10.1063/5.0152374

work page doi:10.1063/5.0152374 2023
[11]

CBMS Regional Conference Series in Mathematics, number 92

Chung, F.R.K.: Spectral Graph Theory. CBMS Regional Conference Series in Mathematics, number 92. AMS, Providence, RI (1997)

work page 1997
[12]

Bronstein, M.M., Bruna, J., Cohen, T., Veliˇ ckovi´ c, P.: Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges (2021)

work page 2021
[13]

Journal of Spectral Theory9(1), 195–226 (2019) https://doi.org/10.4171/ jst/245

Bendikov, A., Cygan, W., Woess, W.: Oscillating heat kernels on ultrametric spaces. Journal of Spectral Theory9(1), 195–226 (2019) https://doi.org/10.4171/ jst/245

work page 2019
[14]

p-Adic Numbers, Ultrametric Analysis and Applications10(1), 1–11 (2018) https://doi.org/10.1134/S2070046618010025

Bendikov, A.: Heat kernels for isotropic-like markov generators on ultrametric spaces: A survey. p-Adic Numbers, Ultrametric Analysis and Applications10(1), 1–11 (2018) https://doi.org/10.1134/S2070046618010025

work page doi:10.1134/s2070046618010025 2018
[15]

Ultrametric pseudodifferential operators and wavelets for the case of non homogeneous measure

Kozyrev, S.V.: Ultrametric pseudodifferential operators and wavelets for the case of non homogeneous measure. arXiv preprint (2005). arXiv:math-ph/0412082v3

work page internal anchor Pith review Pith/arXiv arXiv 2005
[16]

Physica A: Statistical Mechanics and its Applications597, 127221 (2022)

Z´ u˜ niga-Galindo, W.A.: Ultrametric diffusion, rugged energy landscapes and tran- sition networks. Physica A: Statistical Mechanics and its Applications597, 127221 (2022)

work page 2022
[17]

Journal of Sta- tistical Mechanics: Theory and Experiment2025, 113501 (2025) https://doi.org/ 10.1088/1742-5468/ae120f

Mor´ an Ledezma, A.: Time-varying energy landscapes and temperature paths: dynamic transition rates in locally ultrametric complex systems. Journal of Sta- tistical Mechanics: Theory and Experiment2025, 113501 (2025) https://doi.org/ 10.1088/1742-5468/ae120f . Open Access

work page doi:10.1088/1742-5468/ae120f 2025
[18]

Avetisov, V.A., Bikulov, A.K., Zubarev, A.P.: Ultrametric random walk and dynamics of protein molecules. Proc. Steklov Inst. Math.285, 3–25 (2014)

work page 2014
[19]

Dragovich, B., Khrennikov, A.Y., Kozyrev, S.V.e.a.:p-adic mathematical physics: the first 30 years.P-Adic Num Ultrametr Anal Appl9, 87–121 (2017) 42

work page 2017
[20]

Physica A: Statistical Mechanics and its Applications583, 126284 (2021)

Khrennikov, A.: Ultrametric diffusion equation on energy landscape to model disease spread in hierarchic socially clustered population. Physica A: Statistical Mechanics and its Applications583, 126284 (2021)

work page 2021
[21]

Biosystems199, 104288 (2021)

Dragovich, B., Khrennikov, A.Y., Kozyrev, S.V., Misic, N.Z.:p-adic mathematics and theoretical biology. Biosystems199, 104288 (2021)

work page 2021
[22]

Journal of Mathematical Biology92(27) (2026) https://doi.org/10.1007/ s00285-025-02340-8

Fuquen-Tibat´ a, A., Cort´ es-Poza, Y., P´ erez-Buend´ ıa, J.R.: Ap-adic reaction- diffusion model of branching coral growth and calcification dynamics. Journal of Mathematical Biology92(27) (2026) https://doi.org/10.1007/ s00285-025-02340-8

work page 2026
[23]

Molecular Biology and Evolution39(8), 174 (2022) https: //doi.org/10.1093/molbev/msac174

Kumar, S., Suleski, M., Craig, J.M., Kasprowicz, A.E., Sanderford, M., Li, M., Stecher, G., Hedges, S.B.: TimeTree 5: An Expanded Resource for Species Divergence Times. Molecular Biology and Evolution39(8), 174 (2022) https: //doi.org/10.1093/molbev/msac174

work page doi:10.1093/molbev/msac174 2022
[24]

The American Natu- ralist125(1), 1–15 (1985) https://doi.org/10.1086/284325

Felsenstein, J.: Phylogenies and the comparative method. The American Natu- ralist125(1), 1–15 (1985) https://doi.org/10.1086/284325

work page doi:10.1086/284325 1985
[25]

Biometrics62(2), 471– 477 (2006) https://doi.org/10.1111/j.1541-0420.2005.00497.x

Ollier, S., Couteron, P., Chessel, D.: Orthonormal transform to decompose the variance of a life-history trait across a phylogenetic tree. Biometrics62(2), 471– 477 (2006) https://doi.org/10.1111/j.1541-0420.2005.00497.x

work page doi:10.1111/j.1541-0420.2005.00497.x 2006
[26]

Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences479(2278), 20220847 (2023) https://doi.org/ 10.1098/rspa.2022.0847

Gorman, E., Lladser, M.E.: Sparsification of large ultrametric matrices: insights into the microbial tree of life. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences479(2278), 20220847 (2023) https://doi.org/ 10.1098/rspa.2022.0847

work page doi:10.1098/rspa.2022.0847 2023
[27]

Ecology90(9), 2648–2648 (2009) https://doi.org/10

Jones, K.E., Bielby, J., Cardillo, M., Fritz, S.A., O’Dell, J., Orme, C.D.L., Safi, K., Sechrest, W., Boakes, E.H., Carbone, C., Connolly, C., Cutts, M.J., Foster, J.K., Grenyer, R., Habib, M., Plaster, C.A., Price, S.A., Rigby, E.A., Rist, J., Teacher, A., Bininda-Emonds, O.R.P., Gittleman, J.L., Mace, G.M., Purvis, A.: Panthe- ria: a species-level datab...

work page doi:10.1890/08- 2009
[28]

Noh, J.D., Rieger, H.: Random walks on complex networks. Phys. Rev. Lett.92, 118701 (2004) https://doi.org/10.1103/PhysRevLett.92.118701

work page doi:10.1103/physrevlett.92.118701 2004
[29]

In: al., B.B

Angulo, J.: Hierarchical Laplacian and its spectrum in ultrametric image process- ing. In: al., B.B. (ed.) ISMM 2019. LNCS 11564, pp. 29–40. Springer, Switzerland AG (2019)

work page 2019
[30]

Statistics and Computing17(4), 395–416 (2007) 43

Luxburg, U.: A tutorial on spectral clustering. Statistics and Computing17(4), 395–416 (2007) 43

work page 2007
[31]

Peliti, L., Pigolotti, S.: Stochastic Thermodynamics: An Introduction, p. 272. Princeton University Press, ??? (2021)

work page 2021
[32]

North- Holland Personal Library

Kampen, N.G.: Stochastic Processes in Physics and Chemistry, 3rd edn. North- Holland Personal Library. North-Holland, Amsterdam (2007). https://doi.org/ 10.1016/B978-0-444-52965-7.X5000-4 44

work page doi:10.1016/b978-0-444-52965-7.x5000-4 2007