EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Simpler is Not Always Better: Phylodynamic Misspecification and Deep-Learning Corrections

Ruopeng Xie iD, Olivier Gascuel iD, Anna Zhukova iD

Molecular Biology and Evolution · 03.11.2026 · 43 (11)

DOI: 10.1093/molbev/msag268 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Molecular Biology and Evolution, online erschienen am 06.11.2026. Verfasst von Ruopeng Xie und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biology, University of Oxford, Oxford, OX1 3EL, UK. Hauptthemen laut MEDLINE: Deep Learning. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: Marie Skłodowska-Curie Actions, French National Research Agency.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Phylodynamics bridges the gap between epidemiology and pathogen genetic data. The birth–death (BD) model and its extensions serve to infer the average number of secondary infections Re and the infection duration d from time-scaled pathogen phylogenies. Moreover, more complex models add extra parameters, such as the average length of the incubation period, the proportion of superspreaders in the infected population, or the speed-up in detection after contact-tracing. However, these additional parameters come at an important computational cost: While the simplest, BD, model, has a closed-form solution, its extensions do not and require numerical methods for their likelihood computation. This leads to increased computational times and potential numerical errors. Therefore, the BD model remains the favorite researchers’ choice for real dataset analyses, and is often applied even in cases where more complex epidemiological aspects are present. We investigated, using simulations, how model misspecification influences inference of Re and d in the phylodynamic framework. We showed that the use of models not accounting for various epidemiological aspects leads to bias. In particular the simplest, BD, estimator tends to underestimate Re in the presence of super-spreading, incubation or contact-tracing, which might be dangerous from the public health perspective. In contrast, deep-learning-based estimators for complex models, accounting for multiple epidemiological factors, performed well in our simulations both on the data where those factors were present and where they were absent. This motivates the use of complex epidemiologically realistic estimators, whose design has recently become possible thanks to deep learning.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningPhylogenyComputer SimulationHumans

Stichwörter der Autor:innen

deep learningepidemiologyestimation biasmodel misspecificationphylodynamics

Autor:innen und Institutionen

  • Department of Biology, University of Oxford, Oxford, OX1 3EL, UK.
  • Institut de Systématique, Evolution, Biodiversité (ISYEB), URM 7205 CNRS, Muséum National d'Histoire Naturelle, SU, EPHE & UA, Paris, 75005, France.
  • Modelling of Pathogens: Evolution, Adaptation and Spread, Institut Pasteur, Université Paris Cité, Institut de Biologie de l'ENS (IBENS), École Normale Supérieure, CNRS, INSERM, Université PSL, Paris, 75005, France.

ORCID-Profile: Ruopeng Xie, Olivier Gascuel, Anna Zhukova

Förderung

  • Marie Skłodowska-Curie Actions
  • French National Research Agency

Zitieren

APA

Xie, R., Gascuel, O., & Zhukova, A. (2026). Simpler is Not Always Better: Phylodynamic Misspecification and Deep-Learning Corrections. Molecular Biology and Evolution, 43(11). https://doi.org/10.1093/molbev/msag268

Vancouver

Xie R, Gascuel O, Zhukova A. Simpler is Not Always Better: Phylodynamic Misspecification and Deep-Learning Corrections. Molecular Biology and Evolution. 2026;43(11). doi:10.1093/molbev/msag268

BibTeX
@article{Xie2026,
  author = {Xie, Ruopeng and Gascuel, Olivier and Zhukova, Anna},
  title = {Simpler is Not Always Better: Phylodynamic Misspecification and Deep-Learning Corrections},
  journal = {Molecular Biology and Evolution},
  year = {2026},
  volume = {43},
  number = {11},
  doi = {10.1093/molbev/msag268}
}
Kostenlos starten →