Jeanne Tous | Julien Chiquet

Jeanne is ingénieur des ponts eaux et forêts (IPEF). Her doctoral thesis is her first professional experience, I am happy to welcome her in our lab.

Summary

Deciphering the processes that govern the distribution of species has long been a major goal of ecology. It is also critical to predict biodiversity responses to future changes and devise reliable conservation strategies. These processes (dispersal, abiotic factors, biotic conditions) can usually not be directly observed and must instead be inferred from observational data, mainly species community distribution data and environmental descriptors, but their complexity and their intermingled influences make them hard to decipher. Among the most recent community ecology methods designed to study this question, Joint Species Distribution Models (JSDMs) are a statistical approach, usually based on multivariate latent variable models, that simultaneously model the distributions of all the species in a community, in relationship with the effects of environmental covariates, and infer a network of inter-species residual (marginal or partial) correlations. In this thesis, we develop original models that mix graphical models (for association networks) with multivariate latent variable models (to figure distribution data) while integrating different types of variables clustering. We show that the modelling of zero-inflation in community data can be critical to the JSDM-based analysis of species association networks, and we study the relationship between species competition/facilitation interactions and JSDM-inferred correlations.

PhD manucript

Journal papers

2026

bioRxiv

Catching the effects of biotic interactions on community data: partial correlations outperform marginal ones with proper abiotic modelling

Jeanne Tous and Julien Chiquet

2026

DOI

2025

bioRxiv

A JSDM with zero-inflation to increase the ecological relevance of analyses of species distribution data

Jeanne Tous, Julien Chiquet, Amy E. Deacon, and 3 more authors

2025

DOI

2026

CSDA
An integrated method for clustering and association network inference

Jeanne Tous and Julien Chiquet

Computational Statistics & Data Analysis, 2026

Abs DOI arXiv Bib

High dimensional Gaussian graphical models provide a rigorous framework to describe a network of statistical dependencies between variables, such as genes in genomic regulation studies or species in ecology. Penalized methods, including the standard Graphical-Lasso, are well-known approaches to infer the parameters of these models. As the number of variables in the model grow, the network inference and interpretation become more complex. The Normal-Block model is discussed, a model that clusters variables and considers a network at the cluster level. This both adds structure to the network and reduces the number of parameters at stake, thereby easing the inference and interpretation of the underlying network. The approach builds on Graphical-Lasso to add a penalty on the network’s edges and limit the detection of spurious dependencies. A zero-inflated version of the model is also proposed to account for real-world data properties. For the inference procedure, two approaches are introduced, a two-step method based on existing approaches and an original, more rigorous method that simultaneously infers the clustering of variables and the association network between clusters, using a penalized variational Expectation-Maximization approach. An implementation of the model in R, in a package called normalblockr, is available on github (https://github.com/jeannetous/normalblockr). The results of the models in terms of clustering and network inference are presented, using both simulated data and various types of real-world data (proteomics and words occurrences on webpages).
@article{tous2026normalblock, title = {An integrated method for clustering and association network inference}, journal = {Computational Statistics \& Data Analysis}, volume = {219}, pages = {108347}, year = {2026}, issn = {0167-9473}, doi = {https://doi.org/10.1016/j.csda.2026.108347}, author = {Tous, Jeanne and Chiquet, Julien}, keywords = {Gaussian graphical models, Sparse networks, Graphical-Lasso, Variational inference, Clustering}, }

Conferences

2024

Inférence de réseaux d’associations à l’échelle de groupes à partir de données d’abondance avec le modèle PLN-Block

J. Tous and J. Chiquet

In actes des 55\ieme journées françaises de statistique, Bordeaux, 2024