top of page

Funded projects


Sebastian Duchene-Garzon (IP), Evolutionary dynamics of Infectious Diseases group - Call 2021

Team website

The group of Sebastian develop multiscale analysis and various scientific approaches structured around 4 axis:

- Phylodynamic methods: Estimating epidemiological parameters using sequence data, including reproductive number. Two key areas of research here are inferring transmission heterogeneity among and within transmission clusters. An emerging technique is using deep learning for this purpose, which holds promise to be computationally efficient and scalable for very large data sets.

- Evolutionary rates of ancient and modern pathogens: Using molecular clock models to detect lineages that have undergone very rapid evolution. This process appears to determine the emergence of hypermutator lineages in leprosy and variants of concern of sars-cov-2. These methods are also potentially useful to uncover the long-term evolution of pathogens that appear to go prolonged periods of zoonosis or dormancy (e.g. the plague bacterium).

- Understanding sources of information in phylodynamics: Phylodynamic models can use sequence data for parameter inference, but in some cases, notably the birth-death, sequence sampling times are also highly informative. This research direction aims to characterise such information to optimise data sets. For example, combining 90% case counts with 10% sequences to achieve high precision and accuracy in estimates of migration rates and transmission parameters

- Simulating infectious spillover: Zoonotic events are at the core of many major outbreaks. This direction aims to understand tools for simulating such events and understanding their frequency and likelihood of leading to onward transmission in humans or organisms of interest. A key benefit of the simulations is to design wildlife sampling strategies and statistical models to better detect such zoonotic events.


Laura Cantini (IP), Machine learning and integrative genomics group

Team website

Laura Cantini is a G5 junior group leader heading the Machine Learning for Integrative Genomics group. Mathematician by training, Laura works at the interphase of machine learning and genomics. Due to the advent of high-throughput technologies, multiple large-scale quantitative measurements, a.k.a. omics, can be accessed for the same set of biological samples or cells. The focus of Laura’s research activity is to design machine learning methods able to co-analyze the numerous available omics data and translate them into actionable biological knowledge. Laura is a CNRS research associate and a Prairie chair.


Michael White (IP), Infectious Disease Epidemiology and Analytics - call 2020

Team website

An inter-disciplinary team bringing together diverse skill sets spanning mathematical modelling and statistics, molecular and serological assays, and the implementation of field-based epidemiological studies and clinical trials. This group aims to use these tools to develop novel diagnostics for multiple infectious diseases ranging from malaria to SARS-CoV-2.

Some of the key projects are:

- MultiSeroSurv: Algorithms and Multiplex Assays for Integrated Serological Surveillance of Malaria and Neglected Tropical Diseases

- VISPA: P. vivax Serology for Elimination Partnership

- PvSTATEM: P. vivax Serological Testing and Treatment: From a Cluster-Randomised Trial in Ethiopia and Madagascar to a Mobile-Technology Supported Intervention

- Mathematical Modelling of P. vivax Transmission

- High-throughput Multi-pathogen Serological Assays

- Mathematical Modelling of Antibody Kinetics


Camille Berthelot (IP), Comparative functional genomics - call 2020

Team website

The main objectives of this group are to identify the gene networks and non-coding regulatory elements that control the advent of menstruation in primates, and to understand how this genetically inherited trait was acquired in primate genomes during the evolution of the human lineage.


Nicolas Rascovan (IP), Infectious Disease Epidemiology and Analytics - call 2019

Team website

The ultimate goal of this G5 group is to investigate how pathogens emerged, spread and changed at the genomic level over the course of human history, and use this information to better understand the complexity of modern infectious diseases. To do so, It will bring together distinct fields: ancient DNA, phylogenetics and pathogen genomics. Most of our knowledge about infectious diseases derives from studies on modern pathogen strains. The studies of this group will aim at using genomic information from ancient pathogens to reconstruct where and when pathogens were in the past and determine the relationships and specific genomic changes from the strains of past infections and epidemics compared to those of modern times. It will also use ancient pathogen genomes to improve clock rates calculations, which are necessary to estimate divergence times between strains.


Rayan Chikhi (IP), Sequence Bioinformatics - call 2018

Team website

We are a new computational team that researches algorithms for big biological data, such as next-generation sequencing data. Our research roots are close to Computer Science, but the primary goal of the group is to apply research products to bioinformatics and biology. Our biological interests include genomics, metagenomics, pan-genomics, transcriptomics and proteomics. The group develops and implements algorithms and data structures into software tools, and also collaborates with biology groups. Some examples of recent projects are the development of data structures to index large collections of sequencing datasets, methods for improving bacterial genome assemblies with long reads, and genome assemblies (giraffe, gorilla Y, mountain goat). Our ongoing projects include the analysis of variants in Alzheimer’s disease whole-genome sequencing data, the development of algorithms on linked-reads sequencing data, and a search engine for all previously sequenced human RNA-seq experiments.

Etienne Simon-Loriere, Evolutionary genomics of RNA viruses - call 2017

Team website

Emerging infectious diseases represent an increasing threat and burden worldwide, and an improved knowledge of these pathogen is crucial to prepare against the epidemic risk. RNA viruses are over-represented among human pathogens, and their rapid evolution constitutes a challenge for the control of these infectious agents, in addition to likely contributing to their emergence risk. In parallel, there is a wide variation in both animal and human risk and outcome of infection, generally encompassing asymptomatic, to more severe and sometimes lethal cases. This junior group aims to explore the basis of the large differences of sensitivity to infection and severe disease in human from a novel, virological and evolutionary perspective. Specifically, the strategy of this junior group will be to combine phylodynamics studies of natural viral infections to an original in vitro system of evolution that includes the genetic and immune diversity of the host.  Understanding  how  the  various  ecosystem  in  which  a  virus  might  multiply  (here  different human populations), can influence the evolutionary trajectories and composition of viral populations, notably  with  respect  to  pathogenicity  and  transmissibility,  will  open  to  new  models  of  disease emergence and spread, and to functional studies on both the host and virus side.




Raphaël Malak

Supervisors: Hugues Aschard & Rayan Chikhi

Metagenome GWAS of human outcomes.

Genome-wide Association Studies (GWAS) have been central to studying the genetics of complex human phenotypes, and there is now tremendous interest in implementing GWAS-like approaches to assess the role of bacteria genetic in human health outcomes. Building over their respective expertise on statistical genetics and sequence alignement, Drs. Aschard and Chikhi have recently developed an innovative framework that addresses several of the challenges in modelling the complex structure of bacteria genetics. The approach consists in a multi-component linear mixed model applied to unitigs, defined as overlapping DNA words of length k (so-called k-mers). While it demonstrated good performances in multiple settings, the approach still suffers some methodological approximation and remains limited in scope. The objectives of the project are to optimize genetic modelling aspects of the approach, to extend it to multi-species GWAS including metagenomes, and to conduct series of real data analyses of human phenotypes. Analyses will include publicly available bacteria sequencing data, and gut microbiome data of both healthy individuals and inflammatory bowel disease cases.


Juliette Bellengier

Supervisors: Olivier Tenaillon (INSERM), Isabelle Rosinski-Chupin

Emergence of epidemic clones in Escherichia coli, the case of ST38 carbapenem resistant clones.


As the COVID-19 pandemic continues to wreak havoc around the world, it has become increasingly evident that an evolutionary perspective is necessary to comprehend the spread of infectious diseases. Meanwhile, antibiotic resistance in bacterial infections poses a significant public health challenge. Among medically relevant bacterial species like Escherichia coli, years of genomic epidemiology have revealed a dynamic landscape of diversification, with the frequent emergence of epidemic clones. Surprisingly, many of these clones display unique specificities, such as ST38's association with carbapenemase resistance. This existence of epidemic clone specificity is however surprising given the ease with which plasmids and alleles involved in resistance are transferred. One interesting hypothesis to explain those expansion is that the drivers of these lineages successes is not on a single mutation but a combination of interacting mutations. To investigate this intriguing hypothesis, a proposed PhD project aims to employ both experimental and bioinformatic approaches. The project will leverage the expertise of two partners in ST38 epidemiology and recent molecular tools that enable laboratory replication of genetic exchanges found in the wild. The ultimate goal is to unravel the underlying mechanisms that drive the success of epidemic clones and inform new strategies for combating antibiotic-resistant bacteria.


Camille Schneider

Supervisors: Lulla Opatowski, Sylvain Brisse, Anne Cori (Imperial Colllege London)

Developing tools for the analysis of bacterial outbreaks in humans by combining mathematical modelling, Genomics and Epidemiology.


Spread of gram-negative multi-drug resistant bacteria (MDRB) in hospitals is a public health priority globally. While microbial genomics is increasingly available in epidemiological studies and routine surveillance, bacterial transmission routes and its drivers are poorly understood. This is due to the absence of models linking bacterial sequences and epidemiological patterns, impeding systematic reconstruction of transmission chains or identification of outbreak-related sequences. This PhD aims to build integrative mathematical modelling tools for the analysis of nosocomial bacterial outbreaks accounting simultaneously for bacterial evolutionary and ecological dynamics, and environment-to-human or human-to-human transmission.

We will develop multi-layer modelling and statistical frameworks, combining the bacterial microevolution and the strain transmission layers to: (1) Reconstruct transmission chains of hospital outbreaks and estimate key epidemiological and evolutionary parameters; and (2) Provide real-time analysis and identification of strains responsible for bacterial outbreaks through environment contamination. We will analyze different hospital datasets available from longitudinal microbiological and epidemiological patients follow-up.

Overall, the project should bring new scientific, methodological, and public health insights: (1) a better knowledge of the contribution of bacterial genome dynamics, including within-host gene transfers, in hospital transmissions; (2) new methodological frameworks to analyse pathogenic bacteria transmission in hospital; and (3) tools to support control decisions.


Eli Barthome
Supervisors: Thomas Bourgeron & Richard Delorme (APHP)
Contribution of regulatory variants to differences in clinical and brain profiles among autistic individuals
Autism is a highly prevalent and heritable neurodevelopmental condition characterized by impairments in social interactions, and repetitive behaviors/interests. Our current knowledge of the biological pathways associated with autism is mostly based on the identification of rare variants impacting protein structure.
However, non-coding variants are suspected to play a role by modulating gene regulation. Thus, we will 1) precisely identify regulatory variants associated with autism, 2) investigate the genes and biological pathways regulated by such variants and 3) study how they affect brain and clinical profiles of the patients.
Mapping the effect of regulatory variants will improve our understanding of the pathways involved in autism, and allow a better stratification of individuals.
Mohamed Mounib Benimam
Supervisors: Jean-Christophe Olivo-Marin & Lucie Peduto
Computational image analysis of the tumor stroma to understand its role in cancer metastasis.  
Breast cancer is the most frequently occurring malignancy in women, and the 5-year survival rate and therapeutic options drops significantly for patients with metastases. The stromal microenvironment is increasingly recognized as a major factor promoting tumor growth, metastasis and resistance to therapy. The understanding of the stromal microenvironment and its modification require quantitative, reproducible and comprehensive analysis of the tumor stroma as an ecological system.
To address this question, we will leverage the dedicated imaging technics and the cutting-edge image analysis methods to map the stromal microenvironment to model the stromal organization.
This project aims at developing a powerful computational pipeline based on advanced Machine Learning for quantitative image analysis of cellular localization and composition within regions of interest and their relationships to help deciphering the role of stromal cell in breast cancer development and metastasis, with a focus on postpartum breast cancer.
Laura Xénard
Supervisors: Guillaume Duménil, Daria Bonazzi, Jean-Yves Tinevez
Growth and morphogenesis of the diplococcus Neisseria meningitidis under external mechanical and chemical stresses.
Neisseria meningitidis is a diplococcal bacterium of about 1 μm in diameter that can live in the human throat asymptomatically. In some cases, bacteria cross the epithelium and reach the blood circulation leading to severe diseases such as septicaemia and meningitis. Inside vessels, bacteria adhere to the endothelium and proliferate in the form of aggregates in a process known as vascular colonization. However, the impact of such diverse microenvironments on cell shape, size and division remains completely unexplored.
In this PhD project, we aim at exploring meningococcal morphogenesis and robustness in
physiological conditions and in response to extracellular stresses. This work will allow us to identify cell adaptation mechanisms to ensure cell survival during infection, and potentially use cell shape control as a new target for disease treatment.
To achieve this, this interdisciplinary project will be organized around three main axes ranging from
the subcellular to the single cell and multicellular tissue scale. Each axis will feature an experimental
approach coupled to an image analysis pipeline for data quantification. The pipeline will be based on the build-up of an efficient bioimage analysis tool for automated and robust cell tracking and lineage
reconstruction, taylored for Neisseria meningitidis growth.
Alex Westbrook
Supervisors: Julien Mozziconaci (MNHN), Romain Koszul, Pablo Navarro-Gil
Computational design of DNA sequences with deep neural network.

In eukaryotes, the long DNA molecules that constitute chromosomes are wrapped into nucleosomes and further organized in loops and domains in the 3D space by other structural proteins such as cohesins. The 3D genome folding has both structural and gene regulatory roles. An important question in the field is to understand the correspondence between the DNA sequence and the positions of these structuring elements since recurrent variations in the genome can affect its proper 3D folding and play a role in the onset of diseases. To answer this question, we plan here to combine new deep learning based computational tools with synthetic genomic possibilities now offered in the Koszul and Navaro teams. We will first investigate how changes in the DNA sequence can lead to changes in nucleosome and cohesin positions in yeast and mice. We will then go beyond the predictions by designing in silico DNA sequences able to position nucleosomes and cohesins in a controlled manner. The project, which lies at the frontiers between synthetic biology and computer science, would represent one of the first projects exploiting deep learning to design in silico a DNA sequence with desired properties and to test these properties in living organisms.


Andrew Holtz
Supervisors: Hervé BourhyAnna Zhukova

Drivers of Epidemics: Development of Phylogeographic Methods for the Identification of Epidemiological Drivers of Virus Epidemics using SARS-CoV-2 and RABV as templates

This PhD project describes the development of a new Maximum Likelihood (ML) method, analyzing large genomic datasets and epidemiological variables to identify factors responsible for the spread of epidemics. This approach builds on previous work carried out in a Bayesian framework (computationally intensive) and extends it to the ML method via fast algorithms. This approach will elucidate COVID-19 transmission by identifying factors involved in regional and global spread. It will be applied to rabies virus (RABV) epidemics, revealing epidemiological factors that promote RABV emergence which will help prevent outbreaks. Sitting at the cross-section of computational biology, genomic epidemiology, and social science, this project brings together experts in the field of infectious diseases to apply novel phylogeographic tools to impact disease control policy. This project will be supervised by Olivier Gascuel and Hervé Bourhy at Institut Pasteur and will include collaboration with leaders in phylogeography, Guy Baele and Philippe Lemey at KU Leuven, and
public health officials in Morocco, Mali and Côte d’Ivoire. The emergence of new epidemics in our global society makes this tool, capable of integrating large genomic datasets and epidemiological variables, essential and is highly needed and anticipated by modelers and epidemiologists around the world.


Maria Lopopolo

Supervisors: Nicolas RascovanLluis Quintana-Murci
A paleometagenomic perspective to major lifestyle transformations in human populations

The transition into agricultural and sedentary lifestyles had a profound impact on the organization and health of ancient human populations, with consequences that still affect modern societies. Agriculture was adopted in at least five unrelated periods and places over human history, but is still not clear whether the effects were comparable in all these cases. In this project, we will investigate the impact of the agricultural transition in human societies, focusing on pre-Columbian South American populations as case study. To do so, we will sequence human and microbial DNA (including pathogens) recovered from hundreds of archaeological samples of centuries-old farmers and hunter-gatherers. Our main goals is tackling fundamental questions such as the demographic and genetic changes that associated to this revolutionary transformation, and also how it contributed to re-shaping the human microbial makeup and the emergence of infectious diseases. We will implement an interdisciplinary approach that integrates the latest advances in ancient DNA, population genetics and human and microbial genomics, with the complementary input from archaeology and anthropology. And this will be achived by combining the expertises in human population genomics of Quintana-Murci lab, with the experience in microbial (paleo)-(meta)-genomics of Rascovan Lab, two INCEPTION teams placed at Pasteur Institute


Cantin Ortiz

Supervisors: Christoph Schmidt-Hieber & Uwe Maskos         
Modelling memory consolidation impairments in Alzheimer’s disease

Impairment of long-term memory is a hallmark symptom of Alzheimer’s Disease (AD). However, the cellular and circuit basis of this devastating symptom has been poorly explored. Communication between the hippocampus and the prefrontal cortex (PFC) is thought to be critical for memory consolidation, and its dysfunction may underlie cognitive deficits in AD. Building on pilot data, here we propose that interactions between the amyloid beta peptide and the cholinergic neuromodulatory system disrupt inhibitory synaptic transmission, leading to impaired hippocampus-PFC communication and consequently memory deficits in AD. We aim to test this hypothesis by combining the core strengths of our teams: We will use an animal model of AD developed by the Maskos team, where mice express a human variant of the amyloid precursor protein. The Schmidt-Hieber group will contribute novel recording techniques from mice performing memory tasks. These experiments will generate high-dimensional data from various modalities corresponding to different scales of the system. To make sense of it, we will develop novel analysis and computational modelling tools to integrate the imaging, electrophysiological and behavioral data. Our goal is to deliver crucial insight into the fundamental cellular and circuit processes underlying memory impairments in AD, and point towards novel therapeutic possibilities.


Viktoriia Gross

Supervisors: Gregory Batt, Imane El-Meouche (Emmanuel Lemichez Unit), IAME Unit (INSERM/APHP)
Janus – An integrative approach to characterize the two sides of enzyme-mediated antibiotic escape: resistance and tolerance

The non-susceptibility of pathogenic bacteria to antibiotic treatments is a major health problem. Bacteria might escape treatments in two ways: being resistant or being tolerant. Whereas resistant bacteria can multiply in presence of antibiotics, tolerant bacteria can merely survive. Yet, tolerance is increasingly recognized as a major player in treatment failure. An increasing fraction of commensal and pathogenic E coli bacteria express extended-spectrum β-lactamases and/or carbapenemases. These enzymes hydrolyze the antibiotic compound. This adaptation mechanism confers resistance to the individual cell and tolerance to the cell population. Disentangling these two effects, having each different clinical consequences, is not possible using standard approaches. Here, our main objective is to propose an integrated framework for the quantitative characterization of enzyme-mediated resistance and tolerance. We will develop an experimental platform using low-volume bioreactors with on-line absorbance measurements and automated
sampling for the systematic off-line measurement of key indicators such as cell counts, and a model of cell responses in their dynamically-changing environments. Through careful experimental design and appropriate computational methods, models will be calibrated for a range of clinical isolates and antibiotic treatments. This will provide a picture of enzyme-mediated antibiotic escape having unprecedented precision


Vincent Mallet

Supervisors: Michael NilgesJean-Philippe Vert
An integrative approach of Deep Learning for structural and chemical biology research

The current state of imaging techniques produces massive amounts of structural biological data. The structure of compounds plays a key role in structure based drug discovery. However most of the computational usage of this structure neglects the amount of data available and mostly relies on deterministic, physics-based tools.

The success of Machine Learning and Deep Learning in the exploitation of a growing amount of available data in other fields is now established. The main successful applications lie in computer vision and natural language processing, were computers skills now exceed humans’. The goal of this project is to adapt these algorithms for structural 3D data.

Common learning algorithms have limitations that make them less efficient for the data at hand. The underlying properties of a 3D shape must be taken into account for improved efficiency of these methods. The goal of this PhD project is thus to leverage recent advances in this direction and to extend them further for structural biology data. In the meantime, we want to apply such new methods to data used at IP to help the data processing and drug discovery process.


Armin Shoushtarizadeh

Supervisors: Thomas GregorPablo Navarro-Gil

Chromatin and transcription dynamics before, during and after mitosis

The topological reorganization imposed on the chromatin during mitosis leads to a global shutdown of the gene expression. How then is the transcriptional program reestablished after division ? Previous work lacks the coupling between spatial and temporal resolution to assess in real-time the interplay between the transcriptional machinery and the physical properties of the chromatin. The goal of this project is to investigate gene regulation with high spatial and temporal resolution before, during and just after mitosis. Using quantitative imaging we will monitor enhancer-promoter contacts and continuously record transcriptional activity, particularly as cells undergo mitosis. Computational and statistical analyses as well as polymer models will be developed to quantify dynamic interactions and associate them to TF activity, mitotic progression and transcriptional outputs.


Chiara Figazzolo (Pasteur-Paris University International doctoral program)

Supervisors: Marcel Hollenstein

Expansion of the genetic alphabet with artificial metal base pairs

The project focuses on exploring the possibility to chemically expand the genetic code, groundbreaking goal strongly pursued by Synthetic Biology. At first, the natural DNA and RNA nucleotides are to be modified by means of organic synthesis in order to change their chemical properties. Afterwards, new potential base pairs are obtained, exploiting the mediation of metals of transition and coordination chemistry instead of hydrogen bonds. The generation of new pairs is followed by their incomporation in strands of non-modified DNA mediated by natural polymerases to check the compatibility of the artificial bases with natural systems. The creation and possibility to replicate libraries of double strands DNA with artificial elements is finally leading to the generation of aptamers, single strands of DNA mimicking the activity of antibodies and binding to specific targets for diagnostics and therapeutics aims. Specifically, in my project the aim is obtaining aptamenrs for NIHT3T cancer cells. If compared with antibodies, aptamers have the great advantage of being more chemically stable, resistant to higher temperatures and easier to synthesize and manipulate.


Mariana Gonzalez (Pasteur-Paris University International doctoral program)

Supervisors: Arnaud Blondel

Functional Molecular Motions as a Target in Drug Design: Application to Nicotinic Acetylcholine Receptors

Nicotinic acetylcholine receptors (nAChRs), are members of a superfamily of ligand-gated ion channels that regulate fast signal transmission at synapses. In order to achieve this, these receptors go through a number of conformational states, as acetylcholine binds to them after being released from presynaptic neurons. At the resting state, the ion channel remains closed. Upon binding of an agonist, the open channel conformation is stabilized allowing the conduction of ions and inducing depolarization of the cell membrane. Prolonged or repetitive agonist administration stabilizes the channel in a desensitized state with reduced response. nAChRs are potential therapeutic targets for central nervous system disorders such as schizophrenia, Alzheimer's disease, Parkinson’s disease and nicotine addiction. To date, 17 different nAChR subunits have been identified (a1-a10 and b1-b4). The subunit composition of each receptor determines its localization, function, agonist sensitivity as well as channel kinetics. The main goal of this project is to gain a more comprehensive understanding of the structure and molecular mechanisms regulating the function and behavior of nAChRs containing the a5  subunit. This will allow to propose positive allosteric modulators for the α4β2α5 receptor to treat nicotine addiction. The a5 subunit is relevant to treat nicotine addiction since a single nucleotide polymorphism (SNP) (D398N) in this subunit has been found to increase lung cancer and nicotine dependence susceptibility. Large genomewide association studies have found that the maximal response of the protein with the conserved amino acids was two times higher with respect to the variant.


Marie Morel

Supervisors: Olivier GascuelEtienne Simon Loriere
Evolutionary Trajectories of Viruses : Adaptation, Convergence and Dynamics

The combination of large population sizes, high replication rates, short generation time and the error-prone nature of their replication lead to the vast genetic diversity observed for RNA viruses in nature. Within their hosts, these viruses generally exist as a population of mutants with genomic sequences that are genetically related but distinct. These features allow them to quickly adapt to new environments which contribute to their high emergence risk. The aim of this project is to understand how the various environments in which a virus might multiply, can influence its evolutionary trajectories. These evolutionary mechanisms can be studied under two angles: at the sequence and at the population levels. To do so, during my PhD project, I develop new tools to study particular evolutionary mechanisms that result mainly from environmental pressure: convergent and parallel mutations or positive selection. Using existing tools, I also study potential changes in intrahost diversity when viruses are submitted to drug treatment. This work presents the interest of studying basic evolutionary processes of viral evolution through a multiscale approach (population and sequence), with the development of bioinformatics tools, and this for viruses that impose an increasing burden on the public health and economy of many countries.


Robin Chalumeau

Supervisors: Jean-Christophe Olivo-Marin, Jost Eninnga
Next-generation structured illumination microscopy for biological imaging.

Fluorescence microscopy is one of the most used tools in modern experimental biology, but classical microscopes (wide field and confocal) cannot image objects smaller than 200 nanometers because they are limited by the diffraction of light. To overcome this limit, several super-resolution techniques have emerged during the last few decades. One of them, called Structured Illumination Microscopy (SIM), is especially well suited for live imaging on living samples, because it provides wide field images with a resolution of 100 nanometers, at a relatively good framerate without degrading neither fluorophores nor cells. In this method, nine wide field images are acquired with specific illuminations patterns, then an algorithmic reconstruction returns the super-resolved image. Jean-Christophe Olivo-Marin’s Bioimage Analysis unit has been working over the last few years with ESPCI-Paristech and Centrale-Supelec on an alternative reconstruction algorithm that aims to improve the speed of SIM by reducing the number of required images from nine to four. The main axes of the project will be firstly to push this new reconstruction to its limits, to adapt the method to 3D imaging and then to reduce the amount of required data by the implementation of compressive sensing protocols. This new structured illumination microscopy technique, with improved image quality and acquisition time, will be applied to relevant biological questions such as the host-pathogen interaction. In particular, providing new data about the intrusion of pathogenic bacteria like Shigella or Salmonella in human epithelial cells should lead to huge improvements in this field of research.


Jérémy Choin

Supervisors: Lluis Quintana & Antoine Gessain

Population genetic approaches to understand common diseases: adaptation and maladaptation to new environments in Melanesia.

Adopting an evolutionary perspective has become highly complementary to clinical and epidemiological genetic studies, as population genetics can provide new insights into the genetic architecture of human disease. Specifically, the advent of high-throughput sequencing, combined with cutting-edge statistical and mathematical frameworks, provide useful information on the way in which selection removes deleterious mutations from human populations and their potential to adapt to a broad range of climatic, nutritional, and pathogenic environments. Melanesia, a sub-region of Oceania, provides with an excellent model to test important hypotheses in population and medical genomics. Specifically, this project aims to (i) reconstruct the demographic history of Melanesian islanders, (ii) understand how such changes in human demography and environments have affected the efficacy of natural selection to remove deleterious mutations in the human genome and (iii) obtain insight into biological functions having participated in human adaptation and maladaptation, thereby affecting human health. Finally, an integrated epidemiological approach will be used to investigate the genetic basis of a common infection in Melanesia caused by the herpes virus 8 (HHV-8) and to explore the co-evolution of the host and the virus. Together, this study will increase our understanding of how human populations have genetically adapted to the different environments they have encountered, as well as detect events of “maladaptation”, thus increasing knowledge on the genetic architecture of human pathologies.


Jonathan Bastard

Supervisors: Lulla Opatowski (Didier Guillemot Unit,) & Laura Temime (Cnam)
Better understand and control the spread of antibiotics resistant bacteria in livestock and at the human-livestock interface

The main objective of the SARAH project is to study the determinants of antibiotic resistant bacteria dissemination in livestock and at the animal-human interface. We develop a methodological research based on mathematical modelling and the analysis of epidemiological, microbiological and demographic data to:

  • Better understand the role played by the different subpopulations in the temporal spread of the bacteria;

  • Quantify the risks of diffusion of resistant bacteria from one population to another;

  • Determine factors associated with the spread of resistant bacteria;

  • Assess the impact of different control strategies on such diffusion.

Our focus is on several multi-resistant bacteria, in distinct contexts and in different livestock productions:

  • Methicillin resistant Staphylococcus aureus (MRSA) in pig production in France

  • ESBL producing Enterobacteriaceae in calves production in France

  • Colistin resistant and other multidrug-resistant Enterobacteriaceae in pig and chicken productions in Vietnam



Felix Hol (IP), Louis Lambrechts (IP), Jean-Baptiste Masson (IP), Christian Vestergaard (IP)
Merging quantitative imaging and statistical modeling to quantify the impact of dengue virus infection on mosquito blood-feeding behavior BitesGoingViral
Blood-feeding behavior is a central component of a mosquito’s capacity to transmit a pathogen. The
number of hosts a mosquito bites, or the amount of time a mosquito spends probing are parameters that can strongly impact pathogen transmission. A salient variable that may shape such dynamics is the pathogen itself. Infection elicits neuronal, anatomical, and immune responses in the mosquito. Through these multifarious effects, the pathogen may shape the decision-making process of the mosquito, altering the balance between exploration and exploitation. Altered feeding dynamics may strongly impact transmission dynamics, yet despite its relevance for pathogen transmission, it is unclear how dengue virus infection affects the blood-feeding behavior of its dominant vector, Aedes aegypti.
To investigate the vector decision process, we propose an innovative approach leveraging quantitative imaging, statistical modeling, and an engineered skin mimic to create a high-throughput behavioral assay. Combining our expertise in characterizing mosquito-virus interactions, and probabilistic decision-making inference to quantify animal behavior, we will compare the blood-feeding behavior of dengue virus-infected Aedes aegypti with their non-infected counterparts. This approach will provide valuable insights into dengue virus transmission dynamics, and the behavioral mechanisms that make Aedes aegypti such a deadly efficient vector.
Elisabeta Vergu (INRAE), Beatrice Laroche (INRAE), Lulla Opatowski (IP), Didier Guillemot (IP)
Modelling the role of microbiota in the acquisition and transmission of antibiotic resistant pathogenic bacteria. MICROMOD
The human microbiota plays major roles in host infections. It can impact host susceptibility, infection outcome and acquisition of pathogens1. In the past decennia efforts have been made to develop realistic ecological models of pathogens growth within animals or humans.
This project aims at encompassing the within-host and the between-host scales in order to study their interplay in the acquisition and transmission of pathogenic, specifically antibiotic-resistant bacteria3. We will: (1) characterize the community assembly of a microbiota and identify microbial species whose behaviour deviates from the neutral processes, (2) model the within-host dynamics of species, to simulate ecological interactions between commensal microorganisms and pathogenic and/or resistant ones (e.g. S. aureus or E.coli in the nasal and gut microbiota, respectively), (3) summarize the information from the community dynamics taking place in order to assess host pathogen susceptibility and transmission potential, and (4) encompass within- and between-host dynamics, while considering inter-individual heterogeneities, in a single model describing the impact of host-to-host contacts on the diffusion of a pathogen in the population. Our project will rely on epidemiological data and samples collected in several epidemiological cohorts, including nasal samples from the European i-Bird project5, and faecal samples from the Birdy project.


Thomas Bourgeron (IP), Richard Delorme (AP-HP) & Jean-François Deleuze (CEA).

Integrative genetics and brain research on autism/ neurodevelopmental disorders in France

Autism and neurodevelopmental disorders (NDD) concern 2-5% of the population. Several genes and brain regions have been identified, but the mechanisms leading to these conditions remain largely unknown. It is for example not yet understood how common and rare genetic variants collectively impact brain development. At term, our project aims to understand the mechanisms of autism/NDD by integrating knowledge from different fields including psychiatry, genetics and neurobiology. Here, we will gather and standardize all available genetic and phenotypic data on autism/NDD (N>40,000 patients) from virtually all clinical centers in France. It requires a collaboration between experts from hospitals (psychiatrists, geneticists, psychologists and neurologists), Information Technology departments and research institutes. We will collect all relevant clinical information including brain imaging (Task 1) ; standardize and analyze the genetic and brain imaging profiles of the patients (Task 2) and develop a website for data visualization/sharing (Task 3). This national collaboration towards the largest group of patients with autism/NDD in France, combined with other international initiatives, will allow a better detection of low/medium risk factors for autism/NDD and to study their interaction with brain anatomy/function. This will (i) increase the yield of diagnostic; (ii) help to guide clinical care and (iii) characterize patients with similar genetic profiles for clinical trials and personalized treatment strategies.


Guy-Franck Richard &  Charles Baroud                                                                         

Pluridisciplinary characterization of double-strand break repair at single-cell level

Molecular mechanisms involved in early stages of tumorigenesis and transitions to malignant phenotypes, can be understood by studying double-strand break repair (DSBR): Molecular steps leading to cell neoplasia involve the initial loss for a normal cell of the ability to detect and fix damaged DNA, normally recognized by proteins specialized in genome maintenance and integrity, such as checkpoint proteins.
In this project, we wish to characterize, at single cell level, the cellular consequences of a single doublestrand break (DSB). We will work with Saccharomyces cerevisiae, which will be imaged and manipulated in microfluidic devices. We will address the following questions: which factors are responsible for the choice of a given repair pathway? Does this depend on the length of checkpoint activation? How may a damaged cell escape checkpoint activation? What cascade of events will eventually lead to normal cell to become cancerous? This work will build on initial results obtained by a collaboration between the Richard team who have extensive experience with double-strand break repair in Saccharomyces cerevisiae and the Baroud team who are experts in microfluidics and quantitative imaging. The INCEPTION funding will allow us to continue funding the post-doc who obtained these initial results.


Ivo Gomperts BonecaNathalie AulnerSpencer Shorte, Soojin Jang & Christophe Zimmer   Artificial intelligence for antibiotic drugs (AI4AB)

The global health challenge of antibiotic resistance creates a pressing need for new antibacterial drugs. Image based high-throughput screening of chemical libraries has proved effective at identifying new molecules active against bacterial infections. However, it remains a key challenge for drug discovery to identify the molecular targets of these compounds. Here, we propose to develop computational approaches to map drugs onto their molecular targets. For this end, we will perform high-throughput imaging of mutant libraries of the bacteria Pseudomonas aeruginosa and Helicobacter pylori, in various growth media and conditions. We will first train deep neural networks to discriminate mutants or families of genes based on the images alone. Using the learned mappings, we will then analyze images of bacteria exposed to antibacterial
compounds in order to infer -or at least narrow down- their targets. After validating the approach on antibiotics with known targets, we will use it to determine previously unknown drug targets. This project (AI for antibiotics, or AI4AB) will yield original computational methods and new insights into the mechanisms of action of drugs against two major bacterial pathogens.




Chiara Zurzolo Unit (IP), Jean-Baptiste Masson G5 (IP)
Identifying Tunneling Nanotube-like Structures in the Developing Cerebellum

Tunneling nanotubes (TNTs) are thin connections that gained international scientific attention as a novel mechanism of intercellular communication for providing a continuous cytoplasmic bridge between cells. By allowing versatile cell-to-cell transport of cargo (e.g. organelles, viruses, and proteins), TNTs have been associated with a wide range of physiological processes and pathological conditions. Unfortunately, due to a lack of a TNT-specific marker, evidence that these structures exist in vivo is scarce. With most studies failing to establish whether the function of TNTs in vitro translate in complex organisms, there is a pressing need to confirm their existence in tissue before their role in pathology can be addressed. In order to overcome this barrier, this project proposes to combine the use of serial sectioning scanning electron microscopy connectomics, with machine learning for the automatic annotation and identification of TNT-like structures in vivo. Through skeletonization pipelines, electrophysiology, and immunohistological approaches, we plan to characterize these structures and test their relevance in the early post-natal mouse cerebellum. Collectively, the results obtained by this holistic imaging and computational strategy will provide the first structural description of TNT-like structures in tissue. 


Eduardo Rocha Unit (IP/CNRS), Philippe Glaser Unit (IP/CNRS/APHP), Amaury Lambert Guillaume Achaz - SMILE, CIRB (Collège de France)

Path2Resistance: Deciphering evolutionary trajectories to characterize emergence and dissemination of multidrug resistant Escherichia coli

The bacteria from the species Escherichia coli are some of the most frequent causes of infectious diseases in the community and in the hospital. These ubiquitous and versatile pathogens are, by their capacity to disseminate antibiotic resistance in the community, major public health threats. Selection for resistance results from transient repeated rounds of exposure to different antibiotics in humans, animals and external environments. Our preliminary data on E. coli carbapenemase producing isolates suggests that resistance is more frequent in certain genetic contexts. Here, we will integrate public data with a broad private collection of sequenced multi drug resistant isolates to characterize how evolution of the genetic context shapes the emergence of multidrug resistant strains.


Marcel Hollenstein G5 (IP), Ricardo Pellarin Unit (IP/CNRS)
Aptamer-labeling as a strategy to determine the structure of macromolecular assemblies

Electron Microscopy (EM) is invaluable for structural determination since it allows visualization of large macromolecular complexes. However, the resolution of EM maps is often too low for a direct determination of the spatial arrangement of the subunits. To address this issue, we propose an aptamer-based labeling strategy combined with integrative modeling to improve the localization accuracy of individual components of the bacterial type VI secretion system (T6SS), a major virulence factor of many Gram-negative bacteria. The aim of the proposed project is to develop a hybrid pipeline for the generation of protein-specific aptamer-labels that will significantly improve the localization accuracy of individual components within the EM density map of protein complexes.


Sigolène Meilhac Unit (IP/INSERM), Jean-Christophe Olivo-Marin Unit (IP/CNRS)

Computational analysis of 3D cell architecture : application in quantifying myocardium orientation at the cellular and tissue levels

The structure of the cardiac muscle underlies the efficient contraction of the heart. This is not only a question of size, underlying contractile power, but also a question of orientation, determining specific patterns of contraction. We have previously shown that growth of the cardiac muscle is already oriented in the embryonic heart. Advances in the understanding of the mechanism of myocardium orientation have been hampered by an absence of tools and methods to automatise image segmentation and deliver accurate, quantitative descriptions of orientations in 3D. Our objective is to develop segmentation tools to extract orientations at the tissue and cellular levels and identify factors required for oriented myocardial growth. In control mouse embryos, we will test whether cell division orientation correlates with the geometry of cells. Alternatively, using control and manipulated embryos, we will test the effect of three molecular pathways, previously shown to regulate cell division orientation in non-cardiac cell types. This collaborative project combines interdisciplinary expertise in heart development and quantitative image analysis to provide novel insight into the regulation of cardiac muscle growth, with potential applications in tissue engineering for cardiac regenerative medicine.

Sigolène Meilhac, G5 (IP/Imagine), Timothy Wai, G5 (IP), Christophe Zimmer Unit (IP), Francesca Raimondi (APHP/INSERM)

Machine learning to get at the heart of diagnostic cardiology

Congenital heart diseases are devastating developmental disorders that include severe cases characterized by complex anatomical abnormalities. Diagnosis remains a significant challenge that requires high-level expertise in imaging techniques and access to sophisticated equipment. Advances in cardiac biology and medicine have been hampered by an absence of tools and methods able to deliver accurate, quantitative descriptions of cardiac abnormalities at the cellular and organ level. Our objective is to develop deep learning approaches to address these unmet medical needs in cardiology and to accelerate fundamental research in congenital heart diseases by leveraging annotated images in clinical and research archives. From images in patients, mouse models and cellular models, we will address both the anatomical defects and the cellular mechanisms associated with congenital heart diseases. This collaborative project combines interdisciplinary expertise in heart development, cardiac metabolism, paediatric cardiology and image analysis by machine learning to provide novel insight into the origin of heart diseases and to develop novel diagnostic aides to facilitate the dissemination of expertise beyond specialized cardiology reference centres.


Marc Lecuit Unit (IP/INSERM), Lluis Quintana Unit (IP/CNRS), Hugues Aschard, G5 (IP)
Host and bacterial factors involved in invasive listeriosis

Listeriosis is a severe foodborne infection caused by the bacterium Listeria monocytogenes (Lm), a ubiquitous bacterium mainly found in dairy products and processed meat. It manifests as septicemia, central nervous system or maternal-fetal infection. Its overall mortality is very high, above of 25%, and death frequently occurs in patients under antimicrobial therapy. While human exposure to Lm is common, only few patients develop listeriosis, suggesting a role of host and bacterial genetics in this ifnection. Bacterial genomic analyses have identified putative virulence factors that were confirmed experimentally, yet these factors are only partly predictive of the virulence of Lm strains. This project aimed to study the host and bacterial genomes of ~1,000 patients from the MONALISA cohort, in order to identify host and bacterial genetic factors involved in the different forms of listeriosis. Using innovative statistical methods, we perform a human genome-wide association study (GWAS) of Lm infection, a bacterial GWAS of disease severity and clinical presentation, and a genome-to-genome analysis that combines the genomes of both organisms to detect host-pathogen interactions. Our analyses have already identified a genetic host factor that increases by ~3.5 the odds of developing maternal-fetal Lm infection. The identification of host and bacterial risk factors for listeriosis will shed light on the molecular mechanisms underlying listeriosis pathophysiology, and pave the way for preventive measures for at-risk individuals, which may result in reduced disease incidence.

Sylvain BrisseSimon CauchemezAnnick Opinel, Nadia Fernandes (postdoc fellow)

Understanding whooping cough resurgence in Europe by combining genomic, epidemiological and sociological approaches

Whooping cough, caused by the bacterium Bordetella pertussis (Bp), can lead to lethal infections in neonates. Although largely controlled by vaccination, the infection is resurging in several parts of the world, including Europe. To understand the causes of resurgence, it is essential to define B. pertussis population composition and evolutionary changes. This has so far been impossible because of a lack of good genetic data of Bp at the European level. The objectives of the project are to decipher pertussis re-emergence by a population genomics approach complemented by epidemiological modelling and social sciences. We will initiate genomic sequencing within EupertStrain, the European network of national reference centers, leading to the first large-scale genomic sequence dataset of Bp isolates at European scale, and will analyze this unique resource (~2500 genomes and epidemiological surveillance data) to define inter-country dependencies in epidemiological patterns, strain transmission, evolutionary changes and vaccination strategies. We will also investigate historical or sociological factors, such as vaccine hesitancy (leveraging the ResiVax network), that may influence vaccination policy making. The integration of knowledge on pertussis epidemiology and population evolution with public health strategy build-up will contribute to a more efficient answer to the challenges of pertussis resurgence.


Tamara Giles-VernickSean KennedyEtienne Simon-LorièreVictor NaratRomain Duda (postdoc fellow), Guillaume Lachenal (médialab)
Microbial and viral circulations among people and wild and domesticated animals in an ecotone, Democratic Republic of Congo (MICROTONE)

Zoonotic transmissions are a major global health risk, with human-animal contact frequently raised as a driver of emergence. Studies of zoonotic transmission risks are often piecemeal, targeting bushmeat, specific animal reservoirs, or single pathogens. Our study will examine ecological pathways and networks facilitating microbial and viral flows between people and animals and why these flows occur. Our primary objective is to conduct comparative metagenomic analyses of virome and gut microbiome among people and selected wild and domesticated animals along a gradient of ecological change in a forest-savanna mosaic in Democratic Republic of Congo, an epicenter of zoonotic disease emergence. We will analyze potential viral and bacterial overlap among humans and animals and explain this overlap (or not) through analyses of human and animal mobilities, practices and contacts.

We mobilize social sciences, animal ecology, and metagenomics tools to evaluate microbial dynamics among humans, domesticated animals (cows, goats, dogs, poultry) and wild animals (bonobos, other nonhuman primates, bats, rodents, antelopes). This multi-disciplinary, multi-species investigation in an ecotone (a transitional ecological zone linked to zoonotic emergence) will offer a “pre-history” of spillover and emergence, tracing an ecological web of virome and microbial sharing among humans and animals, and elucidating why such flows occur.

bottom of page