Mots-Clés
Gut microbiota
Bayesian modeling
Metapopulation dynamics
Occupancy model
Amplicon sequencing data
Description
Internship title: Modeling the ecological dynamics of host-to-host dispersal in wild mouse gut microbiota
Envisioned starting date: Jan – March 2027 (flexible)
Internship supervisors (to be contacted by candidates): Guilhem Sommeria-Klein (Chargé de Recherche Inria), Aura Raulo (Fellow at Queens College, Oxford U., UK)
Internship summary:
Gut microbes shape many aspects of mammalian biology, from digestion to immune system regulation and nervous system signaling. Yet, the gut microbiota exhibits surprisingly strong variations in composition within populations, for reason that are only partly understood (Huttenhower et al., 2012). In particular, the role of host-to-host dispersal in shaping these variations remains poorly understood. Microbial transmission between host individuals is particularly strong in social species. Among socially interacting hosts, microbiota can be considered to form a metacommunity structured by social networks among hosts (Sarkar et al., 2020). Metacommunity ecology was originally developed to explain how species communities assemble in archipelagos, but it can be adapted to understand ecology of microbiomes nested inside social networks. Following metacommunity theory, microbiome community variation can be modeled as a function of both dispersal (i.e., transmission) and local selection within hosts (e.g., immune filtering).
The goal of this internship is to understand the metacommunity dynamics of gut microbiota in wild wood mice (Apodemus Sylvaticus). These mice maintain strongly structured social networks that are only weakly correlated with individuals’ home ranges. This provides a unique opportunity to disentangle the effect of direct social contact from the effect of exposure to the same environment on microbiome sharing. To this end, an experimental study system has been set up in Wytham Woods near Oxford, UK, in which gut microbiota composition, social interactions and home ranges of hundreds of wild mice are tracked through time through a combination of microchips fitted on mice, forest floor detectors and regular mouse trapping (Raulo et al., 2021). Using this set-up, feces from virtually all individuals (~100) living over a one-hectare forest plot were collected once every one to two weeks over one winter season—a time of the year with stable and uniform seed diet unlikely to drive changes in microbiota¬—for a total of 900 samples. Each sample was subjected to amplicon sequencing to profile gut bacterial communities through 16S rRNA marker gene.
The first objective of the internship will be to develop a probabilistic model of presence-absence dynamics of gut bacteria parameterized by the colonization and extinction rates of each bacterial taxon in each mouse host. Unlike larger organisms, bacteria can undergo rapid clonal growth, whereby their abundance in a system depends primarily on local selection patterns, whereas dispersal defines which bacteria is present in the community in the first place. Following this rationale, the model will encode presence-absence dynamics and abundance dynamics separately. Specifically, the colonization rate will not affect abundances per se, but will be made dependent on the incoming dispersal flux from other mice, which itself depends on the taxon’s abundance in source individuals. Dispersal flux will be additionally weighted by the social interaction network, which describes transmission by direct contacts, and the network of overlapping home ranges, which describes indirect exposure to the same environments. These two networks can be considered as proxies for two different dispersal routes (Raulo et al., 2024). This model will build on the metapopulation modeling framework developed for macro-organisms on islands (Hanski and Ovaskainen, 2003). This model will be applied to the most widespread bacterial taxa to measure their colonization and extinction rates. The second objective will be to expand the model so as to accommodate low-abundance microbes, whose uncertain detection may heavily bias the inference of colonization and extinction rates. Finally, the intern will explore differences in phenotypic traits related to dispersal among microbes, such as aerobic tolerance, spore formation, or proxies of pathogenicity, and investigate how these different traits influence dispersal dynamics and pathways.
Desired skills:
- Master in ecological modeling or computational biology
- Bases in statistics and probabilities
- R or Python programming
- English proficiency
References:
Hanski, I. and Ovaskainen, O. (2003) Metapopulation theory for fragmented landscapes. Theoretical Population Biology 64: 119–127.
Huttenhower, C., Gevers, D., Knight, R., Abubucker, S., Badger, J.H., Chinwalla, A.T., et al. (2012) Structure, function and diversity of the healthy human microbiome. Nature 486: 207–214.
Raulo, A., Allen, B.E., Troitsky, T., Husby, A., Firth, J.A., Coulson, T., and Knowles, S.C.L. (2021) Social networks strongly predict the gut microbiota of wild mice. The ISME Journal 15: 2601–2613.
Raulo, A., Bürkner, P.-C., Finerty, G.E., Dale, J., Hanski, E., English, H.M., et al. (2024) Social and environmental transmission spread different sets of gut microbes in wild mice. Nat Ecol Evol 8: 972–985.
Sarkar, A., Harty, S., Johnson, K.V.-A., Moeller, A.H., Archie, E.A., Schell, L.D., et al. (2020) Microbial transmission in animal social networks and the social microbiome. Nat Ecol Evol 4: 1020–1035.
Sarkar, A., Raulo, A., Stothart, M.R., Ibata, N.G.O., Harty, S., Metcalf, C.J.E., et al. (2026) The island biology of the host microbiome. Trends in Microbiology 34: 627–649.