Pith. sign in

REVIEW 4 major objections 5 minor 78 references

Genotype networks drive oscillating endemicity and epidemic trajectories in viral evolution

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that genotype-network topology alone can decide whether a virus settles into steady endemic levels or recurrent seasonal waves, and that a network model with immune memory reproduces the order in which H3N2 strains emerge.

desk verdict A useful new epidemic model with a nice synthetic result, but the 'topology alone' claim and the H3N2 validation both need more work before I'd trust the strong version. read the letter →

arxiv 2506.03279 v1 pith:ADRNIHXY submitted 2025-06-03 q-bio.PE physics.soc-ph

Santiago Lamata-Otín , Octavian C. Rotita-Ion , Alex Arenas , David Soriano-Paños , Jesús Gómez-Gardeñes This is my paper
classification q-bio.PEphysics.soc-ph MSC 92D3005C82
keywords genotypenetworksantigenicdriftepidemicmodelingseasonalepidemicscross-immunityimmunememoryinfluenzaH3N2networktopology
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 claims that the wiring of a virus's antigenic genotype network—which genotypes sit one mutation away from which—can by itself decide the qualitative epidemic regime: lattice-like antigenic spaces settle into a steady endemic prevalence, while networks built from star-like mutant swarms (clusters of antigenically similar genotypes) produce persistent seasonal waves. The supporting model, SIMS (Susceptible–Infectious–Mutation–Susceptible), couples contagion with dynamic immune memory: each strain's recovery rate rises as hosts are infected by that strain and by antigenically similar neighbors, then wanes, while mutation moves the virus across the network. On the real H3N2 influenza genotype network the model produces variant landscapes with lineage emergence and turnover that resemble surveillance data, and the simulated order in which strains reach their epidemic peaks correlates with the order in which they were first sampled, for the second largest connected component. A sympathetic reader would care because this suggests that both qualitative epidemic rhythm and possibly the emergence order of future variants could be read off genomic network structure alone.

What carries the argument

The load-bearing object is the genotype network $G=(\mathcal{N},\mathcal{L})$, whose nodes are viral amino-acid sequences and whose links join genotypes differing by a single mutation, together with the SIMS differential equations for each strain $i$: the infectious fraction $\rho_i^I(t)$ grows by contagion, decays at a strain-dependent recovery rate $\mu_i(t)$, and diffuses across the network through a term governed by the normalized graph Laplacian $\ell_{ij}$, while each recovery rate is itself a dynamic memory variable that accumulates immune pressure $\alpha\sum_j \rho_j e^{-x_{ij}/\Delta}$ from strain $i$ and cross-reactive neighbors, weighted by an exponential kernel in genetic distance $x_{ij}$, and decays toward the basal rate $\mu_0$ at rate $\gamma$. The Laplacian diffusion term and the exponential cross-immunity kernel with characteristic length $\Delta=3$ are the two mechanisms that translate network structure into epidemic behavior. The control that isolates the mechanism is setting $\alpha=0$: without immune-memory dynamics, the lattice and both star-swarm networks produce macroscopically identical trajectories, which shows that the memory-topology interaction, not the equations alone, carries the regime distinction.

What would settle it

Run the SIMS dynamics on an ensemble of degree-preserving random rewritings of the second-largest H3N2 component—the same configuration-model ensemble used for the structural comparison in Fig. 2c—and measure the peak-time versus sampling-time correlation of Fig. 5: if randomized wiring reproduces the same correlation and the same seasonal-wave pattern, the claim that the specific network topology drives the regime would be falsified. Alternatively, applying the identical protocol to the largest and remaining large components, where unobserved bridging genotypes are more likely, and finding no correlation would falsify the claim that the SIMS model reconstructs real evolutionary trajectories.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the topology of the antigenic genotype network alone, under the SIMS dynamics, determines whether endemic trajectories consist of persistent seasonal waves or converge into steady dynamics. Synthetic networks that concatenate star-like mutant swarms—with long-tailed degree distributions, high modularity, and disassortative mixing—yield sustained oscillations whose outbreak peaks scale with swarm size, whereas homogeneous lattices converge to a steady endemic equilibrium, recovering the well-known low-dimensional result. On the empirical H3N2 genotype network, the same equations generate complex landscapes of co-circulating and successively dominant mutant swarms, and the normalized sequence of simulated strain peak times correlates with the normalized sequence of first-sampling times recorded in genomic surveillance data, for the second largest connected component. On its own terms, the paper establishes that the structure of the antigenic space is a critical determinant of epidemic trajectories and that the SIMS model partially reconstructs the evolutionary history of the H3N2 virus.

Load-bearing premise

The load-bearing premise is that antigenic distance is adequately represented by the number of amino-acid mutations separating sampled genotypes, with immune protection falling off as $e^{-x/\Delta}$ at a single chosen length scale $\Delta=3$, and that each connected component of the sampled network evolves independently from its first-sampled genotype; if unobserved genotypes bridge the components (the paper itself notes in its Discussion that components differ by only two or three mutations) or if antigenic distance is not proportional to mutation count, both the seasonal-wave prediction and the claimed emergence-time correlation could be artifacts of incomplete sampling.

Editorial extensions

If this is right

  • Qualitative epidemic regime becomes a network-structure diagnostic: modular, disassortative genotype networks with star-like mutant swarms point to recurrent seasonal waves, while lattice-like antigenic spaces point to steady endemicity.
  • Genomic surveillance data could feed forecasts, because the simulated order in which strains peak matches the recorded order of first sampling on the H3N2 second largest component, suggesting network position may estimate the emergence order of future haplotypes.
  • The topology-regime link holds specifically under long-lasting immune memory: when immune acquisition and waning act on comparable timescales ($\gamma=\alpha$), trajectories for all three synthetic structures become macroscopically indistinguishable.
  • Mutant swarms act as amplifiers of outbreaks, since the size of a star-like swarm sets the magnitude of its corresponding epidemic peak, offsetting the suppression imposed by cross-immunity.
  • Omitting the infection-history memory term yields unrealistic variant landscapes, with continuous genotype-swarm alternation or full re-emergence of all lineages, so memory and network structure jointly generate the lineage turnover seen in real outbreaks.

Reading between the lines

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

  • If the mechanism generalizes, the incidence patterns of other rapidly drifting viruses—successive SARS-CoV-2 lineages, dengue serotypes, or phage populations—should be predictable from the modularity and degree heterogeneity of their sampled genotype networks; the paper does not test this.
  • The paper compares real and randomized H3N2 networks structurally but never simulates epidemics on degree-preserving random rewritings; if randomized wiring reproduced the same waves and peak ordering, the causal claim would reduce to degree heterogeneity rather than higher-order topology.
  • The emergence-time correlation is shown for the second largest component only; applying the same protocol to the largest and remaining components, where unobserved bridging genotypes are more likely, would directly test whether missing genomic data short-circuit the prediction.
  • Because cross-immunity is encoded by amino-acid Hamming distance with a single fitted length ($\Delta=3$), replacing the kernel with functional antigenic distances measured in the laboratory would test whether the claimed correlation is an artifact of that distance choice.
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 / 5 minor

Summary. The manuscript introduces the Susceptible–Infectious–Mutation–Susceptible (SIMS) model, a mean-field compartmental ODE in which host recovery rates evolve dynamically with infection history and cross-immunity, while viral mutation diffuses along a genotype network. The authors show that on a linear chain the model recovers traveling-wave antigenic drift; on synthetic star-swarm networks it produces periodic waves, whereas a lattice yields a steady endemic state; and on the H3N2 HA genotype network of Williams et al. it produces community-level lineage successions. They claim that the network topology alone controls whether dynamics are seasonal or endemic and that simulated strain peak times correlate with genomic sampling times, allowing estimation of haplotype emergence times.

Significance. Several components are solid and valuable: the model is well posed, it reduces to standard SIS/SIR/SIRS limits, the memory-core formulation improves on earlier Markovian genotype-network models, and the analytic stationary-prevalence and epidemic-threshold derivations in the supplement are useful. The code is publicly available. However, the two headline claims—network topology alone as the causal determinant of epidemic regime, and the ability to estimate real-world emergence times—are not yet supported by the evidence as presented. The paper's significance is therefore conditional on additional control experiments and quantitative reporting.

major comments (4)
  1. [Epidemic trajectories on synthetic genotype networks; Fig. 3] The abstract's claim that network topology alone determines persistent seasonal waves versus steady endemic states is not isolated from degree-sequence effects. The synthetic structures compared in Fig. 3—lattice, homogeneous star concatenation, and heterogeneous star concatenation—differ simultaneously in degree distribution, degree heterogeneity, clustering, and modularity. The degree-preserving configuration-model ensemble in Fig. 2c and Supplementary Fig. 5 is used only for static structural metrics; the SIMS dynamics of Eqs. (7)–(8) are never run on degree-preserving rewires of these synthetic structures. A dynamic control on degree-preserving randomized networks (or rewired star-swarm graphs with the degree sequence fixed) is needed to attribute the oscillatory regime to modular mutant-swarm topology rather than to hub-and-spoke degree heterogeneity. This missing experiment is load-bearing for the central claim.
  2. [Epidemic trajectories on real world genotype networks; Fig. 5] The reported 'high correlation' between normalized peak times (τ^P_i) and sampling times (τ^S_i) is not quantified: no correlation coefficient, p-value, confidence interval, or null model is reported for the density scatterplot in Fig. 5b, and the visual comparison in Fig. 5c does not constitute a quantitative test. The comparison is also partly circular: the H3N2 genotype network is constructed from the same sequences whose first-sampling dates define τ^S_i, the model is initialized at the first-sampled genotype as the wild type, and edges are exactly single-mutation steps (Methods, 'Genotype networks'). Under these conditions some positive rank association between simulated peak order and sampling order is expected even if the model had no true predictive content. A shuffled-date control, a degree-preserving network null, and a proper statistic with uncertainty are required before the abstract's claim that the framework 'allows for estimating the emergence times of various haplotypes' can be assessed.
  3. [Epidemic trajectories on real world genotype networks; Discussion] The emergence-time comparison is made only for the second largest connected component (Fig. 5), while the abstract and Discussion make a general claim about estimating emergence times; no systematic evaluation over the other components (including the largest and seventh, shown in Fig. 4) is provided. Moreover, the Discussion acknowledges that components may be linked by unobserved genotypes differing by two or three point mutations, yet each component is simulated independently. If the sampled network is missing connecting genotypes, the simulated wave propagation from the first-sampled wild type can be an artifact of the sampling process rather than a reconstruction of true antigenic pathways. The manuscript should either restrict the claim to the tested component or perform a sensitivity analysis to missing edges.
  4. [Effective dynamical equations, Eq. (8); Control parameters] The cross-immunity kernel exp(-x_ij/∆) with a single hand-chosen length scale ∆ = 3 is the main coupling between network distance and immunity in all real-network simulations. No empirical validation or systematic sensitivity analysis for ∆ is provided for the H3N2 network; the supplementary variation in Supplementary Fig. 6 is limited to synthetic structures. Because the shape of this kernel controls which genotypes can escape cross-immunity, the oscillation patterns and the peak-time alignment in Figs. 3–5 could be artifacts of this choice if real antigenic distances are not proportional to Hamming distances with this length scale. A sensitivity analysis over ∆ for the real network, or a data-driven calibration of the kernel, should be reported.
minor comments (5)
  1. [Fig. 2 caption] The caption contains a typo: 'Fore example' should read 'For example'.
  2. [Table I, Methods] The table header reads 'N3H2' in two columns; this should be 'H3N2'.
  3. [Methods, Configurational model] The sentence 'we have utilized the function configuration model of the networkx library' should refer to the 'configuration_model' function; as written it sounds like a function named 'configuration model'.
  4. [Fig. 5 and main text] The main text refers to 'Figs. 5.c–d', but the figure caption contains only panels (a), (b), and (c); the reference should be corrected or a panel (d) added.
  5. [Supplement, Eq. (S.8) and surrounding text] The notation for the Laplacian eigenvalue is inconsistent: the text mentions both Λmax and Λmin(L); one consistent symbol for the smallest (zero) Laplacian eigenvalue should be used.

Circularity Check

1 steps flagged · score 6.0 of 10

The Fig. 5 peak-time vs sampling-time correlation is substantially circular because both series inherit the same graph-distance-from-root ordering from the identical H3N2 sequence data used to build the network and to define the target sampling dates.

  1. fitted input called prediction [Results, 'Epidemic trajectories on real world genotype networks' (Fig. 5) and Methods Eqs. (7)-(8)]
    "we record the sequence of peak times of each strain i in the simulation, τ P i , and the day at which the first sample corresponding to the strain was recorded in the genomic surveillance data, τ S i . ... Figure 5.b reports a high correlation between the synthetic and real time series, indicating that the SIMS model can reproduce the evolutionary trajectory of the strains composing the second largest component of the antigenic genotype network of the H3N2 virus."

    The genotype network input and the target sampling dates come from the same H3N2 sequence database: the network is built at the amino acid level and connects genotypes differing by one mutation, while the simulation is initialized with the 'first sampled genotype serving as the wild type.' Under the Laplacian mutation term of Eq. (7), the simulated peak order is essentially the graph-distance ordering from that wild-type node. Real first-sampling dates are similarly ordered by mutation accumulation along the same single-mutation edges. The high correlation in Fig. 5.b therefore largely reflects a shared graph-distance-from-root ranking of the two compared series, rather than an independent confirmation of the SIMS mechanism.

full rationale

The synthetic network result (lattice vs star-swarm concatenations) is not circular: it is a comparison of epidemic trajectories on different input topologies, and the absence of a degree-preserving dynamic control is a confounding-variable concern, not a definitional reduction. The model parameters, including the exponential cross-immunity kernel and Δ=3, are stated assumptions rather than quantities fitted to the predicted output. I found no load-bearing self-citation chain and no imported uniqueness theorem. The Discussion's honest caveat that observed components may be linked by unobserved genotypes separated by two or three mutations weakens the causal interpretation of the real-network simulations, but it is a completeness limitation, not a circular step. The one genuinely circular element is the microscopic 'prediction' of emergence times: the target sampling dates are metadata of the same sampled genotypes used to construct the input network, and the simulated peak order is dominated by diffusion distance from the first-sampled genotype, which is also the dominant ordering of real first-sampling dates. The correlation is therefore substantially forced by construction, giving a partial but real circularity in a central claim.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central results depend on the hand-chosen values of alpha, Delta, Dx, gamma, on the exponential cross-immunity kernel, and on the completeness of the sampled H3N2 network. None of these are derived from first principles or independently calibrated in the paper. This makes the empirical H3N2 comparison more fragile than it appears.

free parameters (6)
  • alpha (immunity acquisition rate) = 0.03
    Set by hand; controls the strength of immune memory. No empirical calibration or sensitivity analysis in the main text.
  • Delta (cross-immunity length) = 3
    Controls how far in genotype space immunity spreads; no serological or antigenic cartography calibration is provided.
  • Dx (mutation rate per day) = 1e-5
    Compromise between empirical H3N2-specific rates in Table I, which span about 4e-6 to 8.6e-4 per day; the value influences wave timing.
  • gamma (waning immunity rate) = 0 in main simulations
    Long-lasting immunity is assumed for the synthetic and H3N2 runs; re-emergence regimes appear only when gamma>0 (Supp. Figs. 3, 8).
  • beta and mu0 (infection and basal recovery rates) = beta=0.3, mu0=0.1 (R0=3)
    Standard parameter choice for R0=3; not varied in main figures, so sensitivity of the topology-driven oscillations to these values is untested.
  • initial infected fraction = 0.01
    Small seed of the wild-type strain in each component; arbitrary but standard.
assumptions (5)
  • domain assumption Genotype network edges equal antigenic distances: one amino acid mutation corresponds to a unit of antigenic distance, and the exponential kernel e^{-x_ij/Delta} captures cross-immunity.
    Invoked in Eq. (8) and in the use of the Williams et al. [27] amino-acid network; no serological cartography data are cited to validate this mapping.
  • domain assumption The observed H3N2 genotype network is complete enough that each connected component can be simulated independently with the first-sampled genotype as wild type.
    The Discussion notes components differ by 2-3 mutations and unobserved genotypes may link them; the simulations ignore inter-component jumps.
  • domain assumption Population-level recovery rates mu_i(t) faithfully summarize host infection history and immune status (mean-field memory).
    Eq. (8) replaces per-host immune histories with a single shared recovery rate per strain; no intra-host diversity or age structure is modeled.
  • domain assumption The population is well-mixed with no contact structure or demographic turnover.
    Stated in Methods: 'we assume a well-mixed population where all individuals are equivalent from a microscopic perspective'.
  • ad hoc to paper Cross-immunity decays exponentially with genetic distance with a single length scale Delta=3.
    The functional form e^{-x_ij/Delta} in Eq. (8) and the value Delta=3 are chosen by hand; no independent evidence or sensitivity analysis anchors them.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Genotype networks drive oscillating endemicity and epidemic trajectories in viral evolution." pith.science (2026). https://pith.science/paper/ADRNIHXY

@misc{pith2026250603279,
  author       = {Pith},
  title        = {Pith review of: Genotype networks drive oscillating endemicity and epidemic trajectories in viral evolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ADRNIHXY}},
  note         = {Machine review of arXiv:2506.03279}
}
read the original abstract

Rapidly evolving viruses use antigenic drift as a key mechanism to evade host immunity and persist in real populations. While traditional models of antigenic drift and epidemic spread rely on low-dimensional antigenic spaces, genomic surveillance data reveal that viral evolution produces complex antigenic genotype networks with hierarchical modular structures. In this study, we present an eco-evolutionary framework in which viral evolution and population immunity dynamics are shaped by the structure of antigenic genotype networks. Using synthetic networks, we demonstrate that network topology alone can drive transitions between stable endemic states and recurrent seasonal epidemics. Furthermore, our results show how the integration of the genotype network of the H3N2 influenza in our model allows for estimating the emergence times of various haplotypes resulting from its evolution. Our findings underscore the critical role of the topology of genotype networks in shaping epidemic behavior and, besides, provide a robust framework for integrating real-world genomic data into predictive epidemic models.

Figures

Figures reproduced from arXiv: 2506.03279 by the authors.

Figure 1
Figure 1. a), creating a feedback loop between mutation and contagion dynamics. In summary, the SIMS model consists of n + 1 com￾partments and is governed by six epidemiological pa￾rameters (see Supplementary Table I), capturing the in￾terplay between contagion, immune dynamics, and mu￾tation in a genetically diverse viral population. The dy￾namics of the SIMS model can be captured by the set of coupled differential Eqs. (7)-… view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. b) the oscillations are regular whereas hetero￾geneous mutant swarms produce variability in both size and shape of the individual outbreak (orange curve in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (5 more)
Figure 6
Figure 6. Figure 6: To round off our analysis on synthetic networks, we now introduce a lattice network in the antigenic spaces, resembling the traditional low-dimensional representa￾tions of the antigenic space. In that case, we observe how the epidemic quickly converges to a steady en￾d…
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 3
Figure 3. Figure 3: As a result, identifying mutant swarms corre [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 3
Figure 3. Figure 3: (a) (b) Supplementary [PITH_FULL_IMAGE:figures/full_fig_p024_3.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

78 extracted references · 75 canonical work pages

  1. [1]

    W. H. O. (WHO), Covid-19 dashboard, accessed: 2024- 11-11, URL https://data.who.int/dashboards/ covid19/cases

    work page 2024

  2. [2]

    Z. Chen, A. S. Azman, X. Chen, J. Zou, Y . Tian, R. Sun, X. Xu, Y . Wu, W. Lu, S. Ge, et al., Nature genetics54, 499 (2022)

    work page 2022

  3. [3]

    Roemer, D

    C. Roemer, D. J. Sheward, R. Hisner, F . Gueli, H. Sakaguchi, N. Frohberg, J. Schoenmakers, K. Sato, ´A. O’Toole, A. Rambaut, et al., Nature microbiology 8, 1952 (2023)

    work page 2023

  4. [4]

    P . V. Markov, M. Ghafari, M. Beer, K. Lythgoe, P . Sim- monds, N. I. Stilianakis, and A. Katzourakis, Nature Re- views Microbiology 21, 361 (2023)

    work page 2023

  5. [5]

    Google Flu Trends, Google flu trends , accessed: 2024-11-11, URL https://www.google.com/ publicdata/explore?ds=z3bsqef7ki44ac_&hl= en&dl=en

    work page 2024

  6. [6]

    R. G. Webster, W. J. Bean, O. T. Gorman, T. M. Cham- bers, and Y . Kawaoka, Microbiological reviews 56, 152 (1992)

    work page 1992

  7. [7]

    M. I. Nelson and E. C. Holmes, Nature reviews genetics 8, 196 (2007)

    work page 2007

  8. [8]

    D. J. Smith, A. S. Lapedes, J. C. De Jong, T. M. Beste- broer, G. F . Rimmelzwaan, A. D. Osterhaus, and R. A. Fouchier, science 305, 371 (2004)

    work page 2004

Show all 78 references
  1. [9]

    L. C. Katzelnick, A. Coello Escoto, A. T. Huang, B. Garcia- Carreras, N. Chowdhury, I. Maljkovic Berry, C. Chavez, P . Buchy, V. Duong, P . Dussart, et al., Science374, 999 (2021)

    2021

  2. [10]

    D. G. Streicker, A. S. Turmelle, M. J. Vonhof, I. V. Kuzmin, G. F . McCracken, and C. E. Rupprecht, Science329, 676 (2010)

    2010

  3. [11]

    A. Moya, E. C. Holmes, and F . Gonz ´alez-Candelas, Na- ture Reviews Microbiology 2, 279 (2004)

    2004

  4. [12]

    Mittal, A

    A. Mittal, A. Khattri, and V. Verma, PLoS Pathogens 18, e1010260 (2022)

    2022

  5. [13]

    Y . Guan, D. Vijaykrishna, J. Bahl, H. Zhu, J. Wang, and G. J. Smith, Protein & cell 1, 9 (2010)

    2010

  6. [14]

    M. G. Guzm ´an, G. Kouri, and S. B. Halstead, The Lancet 355, 1902 (2000)

    2000

  7. [15]

    Soriano-Pa ˜nos, arXiv preprint arXiv:2411.02097 (2024)

    D. Soriano-Pa ˜nos, arXiv preprint arXiv:2411.02097 (2024)

    2024

  8. [16]

    A. M. Carabelli, T. P . Peacock, L. G. Thorne, W. T. Harvey, J. Hughes, S. J. Peacock, W. S. Barclay, T. I. De Silva, G. J. Towers, and D. L. Robertson, Nature Reviews Mi- crobiology 21, 162 (2023)

    2023

  9. [17]

    Tuekprakhon, R

    A. Tuekprakhon, R. Nutalai, A. Dijokaite-Guraliuc, D. Zhou, H. M. Ginn, M. Selvaraj, C. Liu, A. J. Mentzer, P . Supasa, H. M. Duyvesteyn, et al., Cell 185, 2422 (2022)

    2022

  10. [18]

    C. E. Van de Sandt, J. H. Kreijtz, and G. F . Rimmelzwaan, Viruses 4, 1438 (2012)

    2012

  11. [19]

    I. M. Rouzine and G. Rozhnova, PLoS pathogens 14, e1007291 (2018)

    2018

  12. [20]

    J. R. Gog and B. T. Grenfell, Proceedings of the National Academy of Sciences 99, 17209 (2002)

    2002

  13. [21]

    L. Y an, R. A. Neher, and B. I. Shraiman, Elife 8, e44205 (2019). 12

    2019

  14. [22]

    Marchi, M

    J. Marchi, M. L ¨assig, A. M. Walczak, and T. Mora, Pro- ceedings of the National Academy of Sciences 118, e2103398118 (2021)

    2021

  15. [23]

    gisaid.org

    GISAID, Global initiative on sharing all influenza data – gisaid , accessed: 2024-11-11, URL https://www. gisaid.org

    2024

  16. [24]

    Nei and S

    M. Nei and S. Kumar, Molecular evolution and phyloge- netics (Oxford university press, 2000)

    2000

  17. [25]

    A. C. Langedijk, B. Vrancken, R. J. Lebbink, D. Wilkins, E. J. Kelly, E. Baraldi, A. H. Mascare ˜nas de Los Santos, D. M. Danilenko, E. H. Choi, M. A. Palomino, et al., Nature communications 15, 3083 (2024)

    2024

  18. [26]

    Wagner, Proceedings of the Royal Society B: Biological Sciences 281, 20132763 (2014)

    A. Wagner, Proceedings of the Royal Society B: Biological Sciences 281, 20132763 (2014)

    2014

  19. [27]

    B. J. Williams, C. B. Ogbunugafor, B. M. Althouse, and L. H´ebert-Dufresne, PNAS nexus 1, pgac143 (2022)

    2022

  20. [28]

    Wagner, Trends in ecology & evolution 26, 577 (2011)

    A. Wagner, Trends in ecology & evolution 26, 577 (2011)

    2011

  21. [29]

    Manrubia, J

    S. Manrubia, J. A. Cuesta, J. Aguirre, S. E. Ahnert, L. Al- tenberg, A. V. Cano, P . Catal ´an, R. Diaz-Uriarte, S. F . Elena, J. A. Garc´ıa-Mart´ın, et al., Physics of Life Reviews 38, 55 (2021)

    2021

  22. [30]

    D ´abilla and P

    N. D ´abilla and P . T. Dolan, Science Advances 10, eado1693 (2024)

    2024

  23. [31]

    A. R. Vahdati and A. Wagner, BMC evolutionary biology 16, 1 (2016)

    2016

  24. [32]

    Newman, Networks (Oxford university press, 2018)

    M. Newman, Networks (Oxford university press, 2018)

    2018

  25. [33]

    Latora, V

    V. Latora, V. Nicosia, and G. Russo, Complex networks: principles, methods and applications (Cambridge Univer- sity Press, 2017)

    2017

  26. [34]

    L. F . Seoane, H. Secaira-Morocho, E. L´azaro, and S. Man- rubia, arXiv preprint arXiv:2411.07110 (2024)

    2024 arXiv

  27. [35]

    Castillo-Chavez, H

    C. Castillo-Chavez, H. W. Hethcote, V. Andreasen, S. A. Levin, and W. M. Liu, Journal of mathematical biology27, 233 (1989)

    1989

  28. [36]

    Andreasen, J

    V. Andreasen, J. Lin, and S. A. Levin, Journal of mathe- matical biology 35, 825 (1997)

    1997

  29. [37]

    Kamo and A

    M. Kamo and A. Sasaki, Physica D: Nonlinear Phenom- ena 165, 228 (2002)

    2002

  30. [38]

    Ferguson and V

    N. Ferguson and V. Andreasen, in Mathematical ap- proaches for emerging and reemerging infectious dis- eases: models, methods, and theory (Springer, 2002), pp. 157–169

    2002

  31. [39]

    Minayev and N

    P . Minayev and N. Ferguson, Journal of The Royal Society Interface 6, 509 (2009)

    2009

  32. [40]

    A. J. Kucharski, V. Andreasen, and J. R. Gog, Journal of mathematical biology 72, 1 (2016)

    2016

  33. [41]

    B. J. Williams, G. St-Onge, and L. H´ebert-Dufresne, PLoS Computational Biology 17, e1008606 (2021)

    2021

  34. [42]

    Chard `es, A

    V. Chard `es, A. Mazzolini, T. Mora, and A. M. Walczak, Proceedings of the National Academy of Sciences 120, e2307712120 (2023)

    2023

  35. [43]

    F . G. Cobelens, H. van Deutekom, I. W. Draayer-Jansen, A. C. Schepp-Beelen, P . J. van Gerven, R. P . van Kessel, and M. E. Mensen, The Lancet 356, 461 (2000)

    2000

  36. [44]

    S. Hay, D. Smith, and R. Snow, Lancet Infect Dis. 8, 369 (2008)

    2008

  37. [45]

    P . R. Hamid S, Winn A and et al., MMWR Morb Mortal Wkly Rep 72, 355 (2023)

    2023

  38. [46]

    Bonacina, P .-Y

    F . Bonacina, P .-Y . Bo ¨elle, V. Colizza, O. Lopez, M. Thomas, and C. Poletto, medRxiv pp. 2024–08 (2024)

    2024

  39. [47]

    Parino, E

    F . Parino, E. Gustani-Buss, T. Bedford, M. A. Suchard, N. S. Trov ˜ao, A. Rambaut, V. Colizza, C. Poletto, and P . Lemey, PNAS nexus4, pgae561 (2025)

    2025

  40. [48]

    Koelle, S

    K. Koelle, S. Cobey, B. Grenfell, and M. Pascual, Science 314, 1898 (2006)

    2006

  41. [49]

    Pastor-Satorras, C

    R. Pastor-Satorras, C. Castellano, P . Van Mieghem, and A. Vespignani, Reviews of modern physics 87, 925 (2015)

    2015

  42. [50]

    Granell, S

    C. Granell, S. G ´omez, J. G ´omez-Garde˜nes, and A. Are- nas, Annalen der Physik 536, 2400078 (2024)

    2024

  43. [51]

    J. D. Y eakel, J. P . Gibert, T. Gross, P . A. Westley, and J. W. Moore, Philosophical Transactions of the Royal Society B: Biological Sciences 373, 20170018 (2018)

    2018

  44. [52]

    N. Blot, C. Brooks, D. W. Swartz, E. Abdelaleem, M. Garic, A. Iglesias-Ramas, M. Pasek, T. Mora, and A. M. Walczak, arXiv preprint arXiv:2410.17418 (2024)

    2024 arXiv

  45. [53]

    Zhang, Z

    X. Zhang, Z. Ruan, M. Zheng, J. Zhou, S. Boccaletti, and B. Barzel, Nature communications 13, 6218 (2022)

    2022

  46. [54]

    C. A. Smith and B. Ashby, Evolution, Medicine, and Public Health 11, 90 (2023)

    2023

  47. [55]

    Kojaku, L

    S. Kojaku, L. H ´ebert-Dufresne, E. Mones, S. Lehmann, and Y .-Y . Ahn, Nature physics17, 652 (2021)

    2021

  48. [56]

    Lamata-Ot´ın, A

    S. Lamata-Ot´ın, A. Reyna-Lara, and J. G´omez-Garde˜nes, Chaos, Solitons & Fractals 189, 115592 (2024)

    2024

  49. [57]

    Steinegger, L

    B. Steinegger, L. Arola-Fern ´andez, C. Granell, J. G´omez- Garde˜nes, and A. Arenas, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engi- neering Sciences 380, 20210119 (2022)

    2022

  50. [58]

    C. M. Saad-Roy and A. Traulsen, Proceedings of the Na- tional Academy of Sciences 120, e2311584120 (2023)

    2023

  51. [59]

    Traulsen, S

    A. Traulsen, S. A. Levin, and C. M. Saad-Roy, Pro- ceedings of the National Academy of Sciences 120, e2303546120 (2023)

    2023

  52. [60]

    D. H. Morris, F . W. Rossine, J. B. Plotkin, and S. A. Levin, Communications Physics 4, 78 (2021)

    2021

  53. [61]

    Zhang, B

    Y . Zhang, B. D. Aevermann, T. K. Anderson, D. F . Burke, G. Dauphin, Z. Gu, S. He, S. Kumar, C. N. Larsen, A. J. Lee, et al., Nucleic acids research 45, D466 (2017)

    2017

  54. [62]

    M. E. Newman, SIAM review 45, 167 (2003)

    2003

  55. [63]

    Hagberg, P

    A. Hagberg, P . J. Swart, and D. A. Schult, Tech. Rep., Los Alamos National Laboratory (LANL), Los Alamos, NM (United States) (2008)

    2008

  56. [64]

    Euler, Commentarii academiae scientiarum Petropoli- tanae pp

    L. Euler, Commentarii academiae scientiarum Petropoli- tanae pp. 128–140 (1741)

  57. [65]

    E. W. Dijkstra, in Edsger Wybe Dijkstra: his life, work, and legacy (Springer, 2022), pp. 287–290

    2022

  58. [66]

    M. E. Newman, Physical review E 67, 026126 (2003)

    2003

  59. [67]

    J. G. Foster, D. V. Foster, P . Grassberger, and M. Paczuski, Proceedings of the National Academy of Sci- ences 107, 10815 (2010)

    2010

  60. [68]

    Saram ¨aki, M

    J. Saram ¨aki, M. Kivel ¨a, J.-P . Onnela, K. Kaski, and J. Kertesz, Physical Review E 75, 027105 (2007)

    2007

  61. [69]

    M. E. Newman and M. Girvan, Physical review E 69, 026113 (2004)

    2004

  62. [70]

    Rosvall and C

    M. Rosvall and C. T. Bergstrom, Proceedings of the na- tional academy of sciences 105, 1118 (2008)

    2008

  63. [71]

    Edler, A

    D. Edler, A. Holmgren, and M. Rosvall, The MapEqua- tion software package , https://mapequation.org (2025)

    2025

  64. [72]

    Lamata-Ot´ın, A

    S. Lamata-Ot´ın, A. Reyna-Lara, D. Soriano-Pa˜nos, V. La- tora, and J. G ´omez-Garde˜nes, Physical Review E 108, 024305 (2023)

    2023

  65. [73]

    Nobusawa and K

    E. Nobusawa and K. Sato, Journal of virology 80, 3675 (2006)

    2006

  66. [74]

    J. D. Parvin, A. Moscona, W. Pan, J. Leider, and P . Palese, Journal of virology 59, 377 (1986). 13

    1986

  67. [75]

    Sanju ´an, M

    R. Sanju ´an, M. R. Nebot, N. Chirico, L. M. Mansky, and R. Belshaw, Journal of virology 84, 9733 (2010)

    2010

  68. [76]

    M. D. Pauly, M. C. Procario, and A. S. Lauring, Elife 6, e26437 (2017)

    2017

  69. [77]

    N. G. Reich, S. Shrestha, A. A. King, P . Rohani, J. Lessler, S. Kalayanarooj, I.-K. Y oon, R. V. Gibbons, D. S. Burke, and D. A. Cummings, Journal of The Royal Society Inter- face 10, 20130414 (2013)

    2013

  70. [78]

    Tamura, T

    S.-i. Tamura, T. Tanimoto, and T. Kurata, Japanese journal of infectious diseases 58, 195 (2005). 14 SUPPLEMENTARY INFORMATION OFGENOTYPE NETWORKS DRIVE OSCILLATING ENDEMICITY AND EPIDEMIC TRAJECTORIES IN VIRAL EVOLUTION DICTIONARY OF THE PARAMETERS OF THE SIMS MODEL Parameter...

    2005

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.