Mélanie PRAGUE

Bordeaux Population Health, 146 Rue Léo Saignat, 33000 Bordeaux
Phone +33 5 57 57 45 31

Since 2016, I am researcher (CR2) inria at the Bordeaux Sud-Ouest Center, in the team SISTM (common with Inserm in U1219 and Université de Bordeaux).

I got a PhD in Public Health option Biostatistics at the University of Bordeaux in 2013 on the «monitoring of patients infected with HIV». Before that, I‘m a engineer in statistics from ENSAI and have a master degree in mathematical statistics and econometrics. Following my thesis, I conducted a short postdoctoral stay at the University of Oslo (Norway), then I was a postdoctoral fellow for 2 1/2 years at Harvard School of Public Health (Boston, USA).

My research focuses on the development of statistical methods for treatment and prevention of infectious diseases. I draw motivation for the methods from example in sexually transmitted diseases such as HIV. The methods we develop are generic in the sense that they can be applied to other fields than medicine and various medical concerns. During my PhD, I specialized in mechanistic modeling based on Ordinary Differential Equations (ODE) then during my postdoctoral positions we proposed methods to inference a treatment or intervention effect based on causality theory using semi-parametric statistics. In both experiences, once the statistical problem has been modeled, we focus on optimization, to say the treatment delivery and personalization based on control theory and the design or analysis of intervention trials using network sciences.

Key words

  • Mathematics - Statistics and probability: Bayesian; Causality; Control theory; Dynamical model; GEE; Ordinary Differential Equation; Maximum Likelihood; Mediation analysis; Missing data; Network; Semi-parametric.
  • Public Health - Data and statistics: Biomarkers; Biostatistics; High-dimensional data; HIV; Immunology; Personalized medicine; Pharmacometrics; Randomization; Transcriptomic data; Trials design; Vaccine.

Research activities

  • Mechanistic Dynamical modeling
    • Inverse problems
    • System vaccinology
    • In silico trials
  • Evaluation of intervention effect for infectious diseases
    • Causality
    • Missing data
    • Network science
  • Applied expertize
    • HIV/AIDS
    • Ebola virus
    • Nipah virus

Activities linked with research


Actually, amount of hours spent teaching is approximatively 64 hours a year from which half are lectures. The students are from Master in biostatistics and from Engineering degree (ENSAI).
2016 - present
Previously, I was recipient of a teaching assistant grant. Amount of hours spent teaching is approximatively 64 hours a year.
2010 - 2013

List of lectures
List of lectures and tutorials is available below. Specific content can be sent upon request by mail. Classes in bold are still active
  • PhD Level:
    • Coordinator of the "modeling life science" module for the School of Digital Public Health (2017-)
    • Estimation in mechanistic models (2017-)
  • Master Level:
    • Scientific Critical Reading - Engineering degree 3rd year Biostatistics (2018-)
    • Missing data Advanced class - Engineering degree 3rd year Biostatistics (2017-)
    • Introduction to Missing data - Master 2 in Biostatistics (2016-)
    • Design experiment - Master 2 in Pharmacology (2016-2018)
    • Bayesian statistics - Master 1 in stochastic processes (2017-2018)
    • Introduction to research - Engineering degree 3rd year Biostatistics (2011-2013, 2016-2018)
    • Bayesian Statistics - Master 2 in Biostatistics (2011-2013)
    • Inferential Statistics - Master 2 in Biostatistics (2011-2013)
    • Introduction to SAS - Master 1 in Biostatistics (2010)
  • Under-graduate Bachelor Level:
    • Basics in Statistics & probabilities - Bachelor 1 in Social Science(2010-2011)
    • Basics in IT and office software - Bachelor 2 in Medicine (2010-2013)
  • Continued training :
    • Linear regression & Analyze of variance - Technical Degree in Public health (2010-2013, 2016-)
    • Introduction to R - Bordeaux Public Health Summer School (2016-)
    • Bayesian Statistics - Bordeaux Public Health Summer School (2017-)

Selected Collaborations

Victor de Gruttola is a resaearch Professor at Harvard T.H. Chan School of Public Health . Dr. De Gruttola’s research activities focus on developments of statistical methods required for appropriate public health response to the AIDS epidemic.
2014 - current
Alison Hill is a research Fellow and independent principal investigator at Harvard’s Program for Evolutionary Dynamics . Dr. Hill develops mathematical and computational tools to help us better understand, predict, and treat infectious diseases, with a particular focus on human viral infections including HIV/AIDS.
2016 - current
Elizabeth Turner is an Assistant professor at Duke Global Health Institute and the Department of Biostatistics and Bioinformatics . Dr. Turner has experience working in both epidemiological studies and randomized trials across a range of substantive areas in developed world and resource poor settings.
2015 - current
Rui Wang is an Associate Professor of Population Medicine and Director of the Biostatistics Center in the Department of Population Medicine at Harvard Medical School and the Harvard Pilgrim Health Care Institute . Dr. Wang's research interests include the design, monitoring, and analysis of parallel and stepped-wedge cluster randomized trials.
2014 - current

Sebastien Benzecky is researcher at inria Bordeaux . Dr. Benzekry focuses on applied mathematics to cancer research.
2017 - current
Annabelle Collin is assistant professor at Institut Polytechnique de Bordeaux. Dr. Collin research activities are centered around data assimilation using sequential strategies.
2017 - current


Marie Alexandre, co-supervision with R. Thiébaut. Mechanistic modeling and optimization of vaccine response in HIV and Ebola.
2018- current
Laura Villain, collaboration - Analyse et modélisation de l’effet des injections d’interleukine 7 sur les patients infectés par le VIH - Université de Bordeaux - mentor D. Commenges - Presently Data Scientist at Université de Bordeaux.
Chloé Pasin, collaboration Modélisation et optimisation de la réponse vaccinale. Application au VIH et Ebola. - Université de Bordeaux - mentor R. Thiébaut - Presently Postdoc at Columbia University.
Patrick Staples, collaboration 2nd and partially 3rd paper- Networks and Statistical Approaches in Biological Contexts - Harvard T.H. Chan School of public health - mentor JP Onnela - Presently Data Scientist at Mindstrong Health.

Postdoctoral fellows
Irene Balelli Modeling response to Ebola vaccines.
2016- current
Laura Villain, Joint modeling of dynamical biomarkers and high-dimensional data.

Research Engineer
Mélany Durand
2017- current

Master interns
Marie Alexandre - Modèle Vaccinal pour Ebola Dynamique des anticorps et de la réponse cellulaire - Master 1 Centrale Marseille - Presently PHD student.
Jun–Aug 2017
Apr-Aug 2018
Paul Tauzia - Utilisation de la déconvolution cellulaire pour détecter des différences d’expression génique - Master 2 MSS Université de Bordeaux.
Apr-Sept 2017
Augusta Alphonse - Proof of concept for an automated gating tool applied to flow cytometry data from a HIV therapeutic vaccine trial - Master 2 in Vaccinology Creteil.
Apr-Sept 2017
Damien Fossat-Cercler - Dynamical modeling of pharmacoepidemiology in observational studies: optimization of methods for numerical integration - Master 2 in Statistics and ENSAI 3rd year - Presently Data Analyst Senior at PwC France.
Apr-Sept 2013
Ana Jarne - Modeling of T-lymphocytes dynamics after InterLeukine-7 injections in HIV infected patients: introduction of a feedback loop - Master 2 in Biostatistics, Université de Pau - Presently Research engineer at Université de Bordeaux.
Apr-Sept 2012
Sybille Masse - Understanding the Effect of Interleukin-7 with mathematical models - ENSAI 2nd year - Presently Junior biostatistician at Lincoln.
Jun-Sept 2011



  • Villain L., Commenges D., Pasin C., Prague M. and Thiébaut R. : Adaptive protocols based on predictions from a mechanistic model of the effect of IL7 on CD4 counts. Statistics in medicine. early view, in press.
  • Novitsky V., Prague M., Moyo S., Gaolathe T., MmalaneM., Kadima Yankinda E., Chakalisa U., Lebelonyane R., Khan N., Powis K., and others : High HIV-1 RNA among Newly Diagnosed People in Botswana. AIDS research and human retroviruses. early view, in press.


  • Iwuji C., Orne-Gliemann J., Larmarange J., Thiébaut R., Tanser F. ,…, Prague M.,…, Dabis F. and the ANRS 12249 TasP Study Group: Universal test and treat and the HIV epidemic in rural South Africa: a phase 4, open-label, community cluster randomised trial. Lancet HIV, 5(3), Pe116-e125. Access.
  • Prague M., Wang R. and de Grutolla V. : CRTgeeDR: An R package for generalized estimating equations with missing data in cluster randomized trials. The R journal 9 (2) 105-115. Access.
  • Turner E, Prague M., Gallis J., Li F. and Murray D. : Review of recent methodological developments in group-randomized trials: part 2 analysis. American Journal of Public Health, 107(7) 1078-1086 Access.
  • Turner E., Prague M., Gallis J., Li F. and Murray D. : Review of recent methodological developments in group-randomized trials: part 1 design. American Journal of Public Health 107(6), 907-915. Access.
  • Ana J., Commenges D., Prague M., Levy Y., Thiébaut R. and inspire 2/3 study group : Modelling CD4+ T cells dynamics in HIV-infected patients receiving repeated cycles of exogenous interleukin 7. Annals of applied statistics 11(3) 1593-1616. Access.
  • Moyo, S., Mohammed, T., Wirth, K. E., Prague, M., Bennett, K., Holme, M. P., ... and Maphorisa C.N. : Point-of-care Cepheid Xpert HIV-1 viral load test in rural african communities is feasible and reliable. Journal of clinical microbiology, 54(12), 3050-3055. Access.


  • Smith, R., Lee, B. Y., Moustakas, A., Zeigler, A., Prague, M., Santos, R., ... and Comans T. : Population modelling by examples ii. Proceedings of the Summer Computer Simulation Conference (SCSC '16) ACM Digital Library,, 51:1--51:8. Access.
  • Prague M., Gran JM., Thiébaut R., Aalen A. and Commenges D. : Dynamic versus marginal structural models for estimating the effect of HAART on CD4 in observational studies: application to the Aquitaine Cohort study and the Swiss HIV Cohort Study. Biometrics 73(1), 294-304 . Access.
  • Prague M. : Use of dynamical modeling for Optimization of treatment in HIV infected patients. Invited paper in Statistical French Society journal 157(2), 19-38. Access.
  • Prague M., Wang R., Stephens A., Tchetgen Tchetgen E and De Gruttola V. : Accounting for interactions and complex inter‐subject dependency in estimating treatment effect in cluster‐randomized trials with missing outcomes. Biometrics 72(4) 1066-1077. Access
  • Le Grand R., Bosquet N., Dispinseri S., Hopewell N., Gosse L., Desjardins D., Shen X., Tomaras G., Saidi H., Prague M., Barnett S., Thiébaut R., Cope A., Scarlatti G. and Shattock R.J : Superior Efficacy of a Human Immunodeficiency Virus Vaccine Combined with Antiretroviral Prevention in Simian-Human Immunodeficiency Virus-Challenged Nonhuman Primates. Journal of virology 90(11):5315-5328. Access


  • Thiébaut R., Drylewicz J., Prague M., Lacabaratz C., Beq S., Crough T., Sekaly R.P., Lederman M.M., Sereti I., Commenges D. and Levy Y : Quantifying and Predicting the Effect of exogenous Interleukin-7 on CD4+T cells in HIV-1 Infection. PLoS Comput Biol. 2014;10(5):e1003630. Published 2014 May 22, doi:10.1371/journal.pcbi.1003630. Access
  • Thiébaut R., Prague M. and Commenges D. : Modèles mathématiques dynamiques pour la médecine personnalisée. Medecine/Science ITMO Santé publique 30(2) 23-26. Access


  • Prague M., Commenges D. and Thiébaut R. : Dynamical models of biomarkers and clinical progression for personalized medicine: the HIV context. Advanced Drug Delivery Review 65(7) 954–965. Access
  • Prague M., Commenges D., Guedj J., Drylewicz J. and Thiébaut R. : NIMROD: A Program for Inference via Normal Approximation of the Posterior in Models with Random effects based on Ordinary Differential Equations. Computer methods and Programs in Biomedecine 111(2) 447-458. Access


  • Prague M., Commenges D., Drylewicz J. and Thiébaut R. : Treatment monitoring of HIV infected patients based on mechanistic models. Biometrics 68(3) 902–911. Access
  • Philip P, Sagaspe P, Prague M, Tassi P, Capelli A, Bioulac B, Commenges D. and Taillard J. : Acute versus chronic partial sleep deprivation in middle-aged people: differential effect on performance and sleepiness. Sleep 35(7) 997-1002. Access

In revision

  • Staples P., Prague M., De Gruttola V. and Onnela JP. : Leveraging Contact Network Information in Clustered Observational Studies of Contagion Processes. Preprint
  • Staples P., Prague M., Onnela JP. and De Gruttola V. : Size of epidemics in Stochastic block models.
  • Staples P., Prague M., Onnela JP. and De Gruttola V. : Leveraging classical analysis of cluster randomized trials with contact network information in infectious diseases.
  • Turner L. and Prague M. : Inverse probability weighting methods for cluster randomized controlled trial data with missing outcomes: A tutorial.


  • CRTgeeDR: an R package for Doubly Robust Generalized Estimating Equations Estimations in Cluster Randomized Trials with missing data.
    The package allows to estimate the intervention effect in cluster randomized trials (CRTs). It implements a semi-parametric GEE estimator accounting for missing data with Inverse probability weighting (IPW) and for imbalance in covariates with augmentation (AUG). The estimator IPW-AUG-GEE is Doubly robust (DR). Importantly, We exhibited that in CRTs most of the available softwares use an implementation of weights that lead to a bias in estimation if a non-independence working correlation structure is chosen. In CRTgeeDR, we solve this problem by using a different implementation of weighting while keeping the consistency properties of the IPW (and thus of the DR).
    Link to the documentation
    Download CRTgeeDR sources
    Dissemination paper
  • N.I.M.R.O.D. Normal approximation Inference in Models with Random effects based on Ordinary Differential equations
    Models based on ordinary differential equations (ODE) are a widespread tool to describe dynamical system. In biomedical sciences, data within each subject can be sparse but information is often gained from between-subjects variability. This makes natural the use of mixed effect models to estimate population parameters. Although maximum likelihood based approaches are a valuable option, both numerical and identifiability issues favor a Bayesian approach which can incorporate prior knowledge in a flexible way. However the combination of difficulties coming from the ODE system and from the presence of random effects raises a major numerical challenge. A normal approximation of the posterior can be obtained by computing the maximum of the posterior distribution (MAP).
    Link to the documentation and sources
    Dissemination paper
  • marqLevAlg : R package for Marquardt-Levenberg minimization.
    This algorithm provides a numerical solution to the problem of minimizing a function. This is more efficient than the Gauss-Newton-like algorithm when starting from points very far from the final minimum. A new convergence test is implemented (RDM) in addition to the usual stopping criterion : stopping rule is when the gradients are small enough in the parameters metric (GH-1G).
    Link to the documentation
    Download marqLevAlg sources
    Dissemination paper

Positions & Education

Researcher (CR2), inria Bordeaux Sud-Ouest, team SISTM (common with INSERM in U1219 and Université de Bordeaux).
2016 - current
Post-doctoral fellow , Harvard School of Public Health Boston USA, Semi-parametric estimation of treatment effect in clustered randomized trials in presence of missing data: application to HIV prevention in South Africa (Victor de Gruttola).
2014 - 2016
32 months
Invited researcher, University of Oslo, Norway, Comparison of propensity score based methods and dynamical methods for estimation of treatment effect in observational studies (Odd Aalen).
1 month
Postdoctoral fellow, INSERM U121, France, Dynamical mechanistic modeling of HIV’ (Rodolphe Thiébaut).
Oct-Dec 2013
PHD and teaching assistant position, INSERM U121 and Université de Bordeaux, France (Daniel Commenges).
Oct 2010-sep 2013
Master 2 Internship , INSERM U121 and Université de Bordeaux, France, Analysis of Phase II/III HIV Clinical trials (Daniel Commenges).
Apr-Sep 2010
SAS programmer, Averion International, Basel Switzerland,Analysis of a Phase III Clinical Trial in oncology.
May-Aug 2009
Consolidator of data basis, INSERM, Unit 897, Bordeaux France, Data entry in nutrition studies.
Jun-Jul 2008


Qualification for assistant professor CNU 26
Ph.D. in Public Health and Biostatistics with highest honnors , Monitoring of HIV infected patients based on dynamic models, Advisor Daniel Commenges, Biostatistics, École Doctorale SP2 (Political science and public health) within the Université de Bordeaux, France
Master of Science (M.Sc.) in Statistics and Econometrics with high honors (dual curriculum) Department of Mathematics, Université de Rennes 1, France.
Master’s Degree in Engineering in Statistics, (equivalent M.Sc) ENSAI, National School for Statistics and Information Analysis, Rennes France.
Bachelor of Science (B.Sc.) in Economics (dual curriculum) , Department of Economics, Université de Rennes 2, France.
Post-Secondary Preparatory Classes, Lycée Montaigne, Bordeaux, France. University-level courses required in preparation for competitive exams into top universities, engineering, and graduate schools (France’s “Grandes Écoles”). Major : Mathematics and Physics.
High school diploma with high honors, – Lycée V. Louis, Talence, FR. Major: scientific, Specialization: Mathematics.


Active Grants
Mechanistic modeling and optimization of the response to HIV & Ebola vaccines (Prague M.). Frame: inria CORDIS-S.
We propose to combine exploratory analysis, dimension reduction methods for high dimensional data, mathematical dynamical modeling and optimal control theory to propose optimal vaccination strategies in term vector choices and timing of vaccine boosts. This grant aims at funding a 3 years of PhD student.
Oct. 2018-Sept. 2021
Prevac-UP (Yazdanpanah Y.). Role: Work Package leader. Frame: European project RIA EDCTP. PI.
This programme is expected to have a significant impact on Ebola prevention and control and will provide useful information on the effectiveness and the suitability of Ebola vaccine strategies in sub-Saharan Africa. We propose to model the long term antibodies response and the role of baseline parasite status of individuals. This grant aims at funding a 2 years of postdoc, travel and equipment expenses.
Jan. 2018-Jan. 2023
EBOVAC 3 (Rodolphe T.). Role: Collaborator. Frame: European project IMI2
The overall aim of EBOVAC3 is to support an essential part of remaining clinical and manufacturing activities required for licensure in the European Union and the United States of a candidate heterologous prime-boost prophylactic vaccine regimen against Ebola virus disease that is under development at Janssen Vaccines & Prevention B.V. We propose modeling of the long term antibodies response to Ebola vaccination. This grant aims at funding a 4 years of postdoc, travel and equipment expenses.
Jan. 2019-Jan.2024
Nipah Virus Study (Raoul H. & Yuan Z.). Role: Work Package co-PI. Frame: Sino-French Agreement Aviesan.
To raise the challenge caused by Nipah virus we propose to develop a program that shall led to a better understanding of the epidemiology of the virus as well as the associated physiopathology, to develop new tools in the field of diagnosis, treatment and prevention of the infection. This grant aims at funding a 2 years of postdoc, travel and equipment expenses.
Sept. 2018-Aug. 2023
EVALUATES (Rondeau V.). Role: Work Package PI. Frame: INCA project (National Institute for Cancer).
Evaluation of the efficacy of treatments on the evolution of tumor size and other criteria of survival. Development of joint models involving dynamical mechanistic model of tumor growth. This grant aims at funding a 6 months of intern students, travel and equipment expenses.
Aug. 2017-Jun. 2022
Revealing Reservoirs during Rebound (Smith D.). Role: Collaborator . Frame: P01AI131385 NIH.. The mission of this research focus will be to create and apply new methods to address analytical challenges arising from HIV reservoir measures and cure issues in anatomic compartments using data from a variety of technological platforms.It supports my travel expenses for USA.
Aug. 2017-Jun. 2022
Previous Grants
Conference Travel Grant (Statistical French Society) . This award allow young scientist whose application get selected to attend a French conference (600€). It allowed me to attend the “Journée de la Statistique 2016”.
Philippe foundation grant (renewal). My project concerned: "Evaluating and targeting of HIV prevention strategies" (6000$ personal grant for research travels).
Philippe foundation grant. It's a non-profit organization for Franco-American exchanges. My project concerned: "Methods for analyzing HIV clustered randomized data and cohort data with informative missingness" (5000$ personal grant).

Awards & Honnors

Prime d'Excellence. It is a special mention delivered by inria at the national level.
Harvard Rose fellowship Award. Four prices a year to offer a cross-cultural experience of research in developing country. My project concerned: "Estimation of incidence and prevalence of HIV in Botswana: pooling data from different registers to evaluate the impact of personal characteristics such as citizenship".
Marie-Jeanne et Laurent Duhamel Award. The price is given once every other year by the Statistical French Society and awards scientific quality of PhD works in Applied Statistics.
Selected for summer school Sidaction Young researchers. One week intensive class program to mix multidisciplinary early-stage researchers in order to promote collaborations in the field of HIV.