I'm a postdoctoral researcher in evolutionary epidemiology at the CEFE, Montpellier, France. I'm currently working with Sylvain Gandon on pathogens immune escape in response to vaccine-induced immunity.
I did my PhD in the ETE modelling team within the MIVEGEC lab, where I was fainly supervised by Samuel Alizon, Mircea T. Sofonea and Ramsès Djidjou-Demasse. It was followed by a first postdoctoral position at the Bordeaux Population Health centre, working with Mélanie Prague on HBV within-host modelling.

reyne.bastien_[AT]_gmail.com

Montpellier
France

I'm broadly interested in thematics linked to epidemiology, ecology and evolution, with some particular interests listed hereafter.

Thematics

Modelling imperfect immunity Evolutionary epidemiology

Modelling imperfect immunity

×

Since the early days of compartmental models, the assumption of perfect immunity has been challenged in favour of more realistic paradigms. While studying endemicity, Kermack & McKendrick, [1932, 1933] developed a framework in which recovered individuals could be directly reinfected \(\left(S\longrightarrow I\longrightarrow R \longrightarrow I\right)\). This is now known as leaky immunity, in contrast to other frameworks such as the polarised one (e.g. the \(\mathit{SIRS}\) model), where individuals are either fully susceptible or completely immune. These immunity paradigms may complement others (e.g. the \(\mathit{SIS}\) model) or be viewed as components of a more generalised framework [Park et al., 2023]. Incorporating the age of immunity into models could also help clarify these differences [Reyné et al., 2025].

Related references

• Leaky or polarised immunity: non-Markovian modelling highlights the impact of immune memory assumptions
• Mutant emergence timing and population immunisation status impact epidemiological dynamics

Evolutionary epidemiology

×

The ability of a new strain/mutant that has just appeared to thrive depends on many factors. These include stochastic events upon arrival, the composition of population immunity, or the epidemiological regime of the resident strain. These different components interact with each other, producing an environmental feedback where the evolution component alters the epidemiological landscape, making the mutant's success non-obvious. In addition to this, considerations on which metric to assess a mutant's success (probability of non-extinction, producing the next epidemic wave, or having a higher growth rate) may lead to antagonist results.

Related references

• Mutant emergence timing and population immunisation status impact epidemiological dynamics

Methods

Non-Markovian compartmental models Differential equations in epidemiology

Non-Markovian compartmental models

×

Most simple classical ODEs-based compartmental models assume constant leaving rates, which induce exponential waiting times that are often unrealistic from the biological point of view [Reyné et al., 2022]. My work mainly explored alternatives to the ODE paradigm, to allow the explicit implementation of biological processes that depend on the time elapsed since an event (e.g. an infection). It may involve partial differential equations [Kermack & McKendrick, 1932, Reyné et al., 2022] or discrete-time models [Sofonea et al., 2021]. These formalisms are very convenient when accounting for long-term biological processes, such as modelling immunity [Reyné et al., 2024].

Related references

• Principles of mathematical epidemiology and compartmental modelling application to COVID-19
• Non-Markovian modelling highlights the importance of age structure on Covid-19 epidemiological dynamics
• Memory is key in capturing COVID-19 epidemiological dynamics

Differential equations in epidemiology

×

All compartmental models in epidemiology rely on a set of differential equations. Classically, it involves ordinary differential equations (ODEs) as they are easy to compute and sufficiently flexible to tackle a wide range of questions. However, at times, some may prefer the use of partial differential equations (PDEs) to model easily time-dependent phenomena (e.g., immunity, infectiousness) that need to be explored more precisely, or discrete-time models that may reproduce the same time discretisation as the released data.

Related references

• Leaky or polarised immunity: non-Markovian modelling highlights the impact of immune memory assumptions
• Epidemiological monitoring and control perspectives: application of a parsimonious modelling framework to the COVID-19 dynamics in France
• Principles of mathematical epidemiology and compartmental modelling application to COVID-19

Pathogens

SARS-CoV-2 HPV

SARS-CoV-2

×

During the first years of the SARS-CoV-2 pandemic, I worked on various epidemiological aspects as a member of the ETE modelling team. The aims and questions changed all along the pandemic, starting with practical needs such as the estimation of interventions in aged-care facilities [Reyné et al., 2021], estimation of the temporal reproduction number, projections of hospital occupancy [Sofonea et al., 2021] or mitigation strategies [Sofonea et al., 2020]. The later projects, more theoretical, wondered the best way to model SARS-CoV-2 immunity or what could lead to the appearance of new strains [Reyné et al., 2025].

Related references

• Epidemiological monitoring and control perspectives: application of a parsimonious modelling framework to the COVID-19 dynamics in France
• Analysing different exposures identifies that wearing masks and establishing COVID-19 areas reduce secondary-attack risk in aged-care facilities
• Memory is key in capturing COVID-19 epidemiological dynamics
• Mutant emergence timing and population immunisation status impact epidemiological dynamics

HPV

×

All my work on HPV has been carried out while I helped people work on the PAPCLEAR clinical cohort (part of the EVOLPROOF project). The main idea behind this project is to focus on acute HPV infections and understand why they do not persist and are cleared within few months [Alizon et al., 2017].
The results from the PAPCLEAR cohort tends to display a cross-protective effect of HPV vaccines on the HPV51 genotype [Murall et al., 2020] and potential link between menstrual cups usage and genital fungal infections [Tessandier et al., 2023]. At a within-host scale, the HPV infection median time is about 14 months [Tessandier, Elie et al., 2025].

Related references

• HPV cervical infections and serological status in vaccinated and unvaccinated women
• Does exposure to different menstrual products affect the vaginal environment?
• Viral and immune dynamics of genital human papillomavirus infections in young women with high temporal resolution

Tools
Modelling: ODEs, PDEs, discrete-time models, NLMEM, GLMs
Programming: R, Rcpp, Stan

A complete list of publications is also available on my ORCID or my Google Scholar profile.

Some random info linked to my professional activities.

I reviewed some scientific articles for journals such as Proceedings B, Biological Letters or J Roy Soc Interface.
The complete list is available on my ORCID profile.

— Doctoral fellowship from University of Montpellier.
— ANRS mobility grant for young researchers.

I attended international and national conferences where I had the opportunity to give some talks (e.g. COVID-19 Dynamics & Evolution Webinar Series, 2020) or present some posters (e.g. Dynamics & Evolution of Human Viruses, 2022).

I used to teach (in French) statistics, biomathematics and modelling for Bachelor's and MSc's students.

I participated in scientific dissemination program where I had the opportunity explain the fundamental aspects of mathematical modelling, as well as discussing vaccination campaigns through game theory approaches.

I was involved in surveillance activities (as member of the ETE modelling team) during the SARS-CoV-2 epidemic. I developed in that context Shiny apps to monitor the temporal reproduction number (still online apparently) or prospective hospital admissions in the short-term (unmaintained).

I studied mathematics (Bachelor's degree) and statistics (MSc) during my training in Montpellier.