Overcoming Model Misspecification in Bayesian Inference of Molecular Signalling Networks

Abstract

Bayesian inference of molecular signalling networks usually relies on tractability of the marginal likelihood, enabling the set of possible networks to be efficiently explored. As such, linear models with independent errors and conjugate priors are routinely used. However, the dynamics of molecular signalling are nonlinear, and relevant confounders are often unobserved; failure to account for these complexities will almost certainly lead to over-confident inferences in the standard Bayesian framework. To confront this reality, we develop a post-Bayesian approach to inference of molecular signalling networks, guided by the principle that uncertainty should not vanish when the statistical model is misspecified, even in the infinite-data limit. Technically, we extend the predictively-oriented (PrO) posterior of McLatchie et al. (2025) to the setting of latent variable models, empirically investigating the properties of PrO posteriors in the challenging network inference context.

Bayesian inference under model misspecification
Jack Jewson
arXiv8 Oct 2026PreprintCC BY 4.0Full textStatisticsarXiv
When Noise Estimation Hides Basis Misspecification in Repeated Bayesian Inverse Problems
Rares Dimitrie Grozavescu, Mark Girolami
arXiv29 Sep 2026PreprintCC BY 4.0Full textStatisticsarXiv
Structure-Informed Bayesian Inference of Anomalous Transport and Hidden Molecular Trapping in Amorphous Media
Andrey Ananev, Maria Potapova, Nikolay Kondratyuk +2
arXiv8 Sep 2026PreprintarXiv licenceFull textPhysical chemistry StatisticsarXiv
Scalable Diffusion SBI for Compositional Inference under Simulator Misspecification
Vincent D. Zaballa, Elliot E. Hui
arXiv29 Sep 2026PreprintCC BY 4.0Full textBioinformatics & computational biology StatisticsarXiv
Evolutionary stochastic gradient Langevin dynamics: overcoming local traps in large-scale Bayesian sampling.
Yushu Huang, Faming Liang
Statistics and computing5 Nov 2026Peer-reviewedCC BYFull textStatisticsEurope PMC

Cite

BibTeX RIS

BibTeX

@misc{li2026overcoming,
  title = {{Overcoming Model Misspecification in Bayesian Inference of Molecular Signalling Networks}},
  author = {Yinghan Li and Heishiro Kanagawa and Francesca Romana Crucinio and Matthew A. Fisher and Chris. J. Oates},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.24590},
  eprint = {2609.24590},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.24590},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Overcoming Model Misspecification in Bayesian Inference of Molecular Signalling Networks
AU  - Yinghan Li
AU  - Heishiro Kanagawa
AU  - Francesca Romana Crucinio
AU  - Matthew A. Fisher
AU  - Chris. J. Oates
PY  - 2026
DA  - 2026/09/21
PB  - arXiv
DO  - 10.48550/arxiv.2609.24590
UR  - https://arxiv.org/abs/2609.24590
AB  - Bayesian inference of molecular signalling networks usually relies on tractability of the marginal likelihood, enabling the set of possible networks to be efficiently explored. As such, linear models with independent errors and conjugate priors are routinely used. However, the dynamics of molecular signalling are nonlinear, and relevant confounders are often unobserved; failure to account for these complexities will almost certainly lead to over-confident inferences in the standard Bayesian framework. To confront this reality, we develop a post-Bayesian approach to inference of molecular signalling networks, guided by the principle that uncertainty should not vanish when the statistical model is misspecified, even in the infinite-data limit. Technically, we extend the predictively-oriented (PrO) posterior of McLatchie et al. (2025) to the setting of latent variable models, empirically investigating the properties of PrO posteriors in the challenging network inference context.
ER  -

Abstract

Related

Cite