Hub members Have many expertise, covering most of the fields in bioinformatics and biostatistics. You'll find below a non-exhaustive list of these expertise
Searched keyword : Machine learning
Related people (10)
CV Senior Bioinformatician August 2015 – Present : Institut Pasteur, Paris PostDoc fellow 2011 – 2015 : Pascale Cossart’s laboratory, Unité des Interactions Bactéries-Cellules, Institut Pasteur, Paris Phd fellow 2007 – 2010 : Institut des Hautes Etudes Scientifiques, ann Ecole Normale Supérieure, Paris Magister of Science, Theoretical Physics 2003 – 2007 : Dynamical systems and statistics of complex matter, Université Paris 7 and Université Paris 6
BiophysicsMachine learningModelingProteomicsBiostatisticsDatabases and ontologiesHost-pathogen interactions
- Analysis of DNA methylation in the presence and absence of antibiotics in wt and mutant V. cholerae(Baharoglu ZEYNEP - Bacterial Genome Plasticity) - Closed
- Finding and Predicting CRISPR-Cas9 Efficiency(Jerome WONG NG - Synthetic Biology) - Closed
- Characterization of a Salmonella mutant carrying a single amino-acid substitution in the stress sigma factor RpoS(Françoise NOREL - Biochemistry of Macromolecular Interactions) - Closed
I received a Ph.D. in Biostatistics and Bioinformatics applied to Cancer Research in 2011 from the University Paris Sud XI, I was working at the Curie institute under the supervision of Emmanuel Barillot and François Radvanyi. My Ph.D. was about the unsupervised analysis of cancer transcriptome. During my postdoctoral time, I worked on the computational and statistical analysis of NGS data. My areas of interest and expertise include - functional genomics - human genetics - statistical analysis of high-dimensional data - normalization, batch-correction, meta-analysis of high-throughput data - unsupervised learning, independent component analysis - NGS data analysis (RNA-Seq, DNA-Seq, …) - analysis of the non-coding genome, transposable elements
- Left-right patterning of heart precursors(Tobias BØNNELYKKE - Heart Morphogenesis) - In Progress
- Collaboration between CETEA, C2RA and Hub for optimization of experimental designs (3R)(Myriam MATTEI - Center for Animal Resources and Research) - In Progress
- Mechanisms of HIV-1-infected cells susceptibility to Fc effector functions(Timothée BRUEL - Virus and Immunity) - In Progress
One of my projects consists in developing GRAVITY, a java tool based on Cytoscape to integrate genetic variants within protein-protein interaction networks to allow the visual and statistical interpretation of next-generation sequencing data, ultimately helping geneticists and clinicians to identify causal variants and better diagnose their patients. I’m also involved in several other projects in the lab, taking part in the design of pipelines for the processing and the analysis of genomics data, including SNP arrays, whole-exome and whole-genome sequencing data. This means being confronted to the big data problematic, the unit having to manage hundreds of terabytes of genomics data. Finally, I am now analysing these data in order to identify possible causes for autism, to help clinicians with their diagnosis but also to better understand the biological mechanisms at play in this complex disease. This is done through the project aiming at understanding the genetic architecture of autism in the Faroe Islands, and also with the newly starting IMI2 European project AIMS2-Trials.
AlgorithmicsData managementData VisualizationGenomicsMachine learningProteomicsGenome analysisBiostatisticsProgram developmentScientific computingApplication of mathematics in sciencesExploratory data analysisSofware development and engineeringData and text miningGenetics
I joined the Bioinformatics and Biostatistics Hub at Institut Pasteur in 2016 where I am currently developing pipelines related to NGS for the Biomics Pôle. I have an interdisciplinary research experience: after a PhD in Astronomy (gravitational wave data analysis), I joined several research institute to work in the fields of plant modelling (INRIA, Montpellier, 2008-2011), System Biology — in particular logical modelling (EMBL-EBI Cambridge, U.K., 2011-2015), and drug discovery (Sanger Institute, Cambridge, U.K.), 2015). On a daily basis, I use data analysis and machine learning techniques within high-quality software to tackle scientific problems.
AlgorithmicsData managementData VisualizationGenome assemblyGenomicsMachine learningModelingScientific computingDatabases and ontologiesSofware development and engineeringData and text miningIllumina HiSeqGraph theory and analysisIllumina MiSeq
Data VisualizationMachine learningStatistical inferenceBiostatisticsApplication of mathematics in sciencesDimensional reductionMultidimensional data analysis
- High content screening of mitochondrial morphology defects in mitochondrial genetic diseases: discovery of new therapeutic compounds(Claire PUJOL - Mitochondrial Biology) - New
- Red Cross health workers and the continuity of care in the COVID-19 pandemic: A mixed methods operational approach(Leonard HEYERDAHL - Group: Medical anthropology and environment) - New
- A genome-wide RNAi screening for mitochondrial fission factors(Erwan RIVIÈRE - Mitochondrial Biology) - Closed
Bernd Jagla received his PhD in bioinformatics (department of Biology, Chemistry, and Parmacy) from the Free University in Berlin, Germany in 1999. Before joining the Institut Pasteur, he worked for almost ten years in New York City, including as an associate research scientist in the Joint Centers for System Biology (Columbia University) and at the Columbia University Screening Center led by Dr J.E. Rothman. He joined the Institut Pasteur in 2009 to take charge of the bioinformatic needs at the Transcriptome et Epigenome platform, focusing on Next Generation Sequencing. As of 2016 he is member of the C3BI – HUB Team detached to the Human immunology center (CIH) and provides support for cytometry, next generation sequencing, and microarray data analysis. His areas of interest include the quality assurance and data analysis and visualization at the facility. He also has strong expertise in developing algorithms for function prediction from sequence data, image analysis, analysis of mass spectrometry data, workflow management systems. While at Pasteur he developed: KNIME extensions for Next Generation Sequencing (Link) Post Alignment Visualization and Characterization of High-Throughput Sequencing Experiments (Link) Post Alignment statistics of Illumina reads (Link)
AlgorithmicsChIP-seqData managementData VisualizationImage analysisMachine learningSequence analysisDatabaseGenome analysisBiostatisticsProgram developmentScientific computingData and text miningIllumina HiSeqGraphics and Image ProcessingIllumina MiSeqHigh Throughput ScreeningFlow cytometry/cell sortingPac Bio
- Identifying new population(s) of NK cells involved in memory to bacterial infection(Melanie HAMON - Chromatin and Infection) - In Progress
- A long-term mission for an assigned CIH-embedded bioinformatician to provide bioinformatic support to the CIH community(Milena HASAN - Department of Immunology) - In Progress
Data managementMachine learningStatistical inferenceScientific computingExploratory data analysisSofware development and engineeringParallel computingNeuroimaging and computational neuroscienceGrid and cloud computing
Dr. Natalia Pietrosemoli is an Engineer with a M. Sc. in Modeling and Simulation of Complex Realities from the International Center for Theoretical Physics, ICTP and the International School of Advanced Studies, SISSA (Triest, Italy). During her M. Sc. internships she mostly worked in modeling, optimization, combinatorics and information theory applied to medical imaging. In 2012 she got a Ph. D in Computational Biology from the School of Bioengineering of Rice University (Houston, TX, US), where she specialized in computational structural biology and functional genomics. Her doctoral thesis “Protein functional features extracted with from primary sequences : a focus on disordered regions”, contributed to a better understanding of the functional and evolutionary role of intrinsic disorder in protein plasticity, complexity and adaptation to stress conditions. As part of her Ph. D., Natalia was a visiting scholar in two labs in Madrid: the Structural Computational Biology Group at the Spanish National Cancer Research Centre (CNIO), where she mainly worked in sequence analysis and the functional-structural relationships of proteins, and the Computational Systems Biology Group at the Spanish National Centre for Biotechnology (CNB-CSIC ), where she studied the functional implications of intrinsically disordered proteins at the genomic level for several organisms, collaborating with different experimental and theoretical groups. In 2013, she joined the Swiss Institute of Bioinformatics as a postdoctoral fellow in the Bioinformactics Core Facility. Her main project consisted in the molecular classification of a rare type of lymphoma, which involved the integration of transcriptomic, clinical and mutational data for the identification of molecular markers for classification, diagnosis and prognosis. This work was performed in collaboration with the Pathology Institute at the University Hospital of Lausanne (CHUV). In November of 2015 Natalia joined the Hub Team @ Pasteur C3BI as a Senior Bioinformatician. Natalia is especially interested in the integrative analysis of different omics data, both at large-scale and for small datasets, and loves collaborating in interdisciplinary environments and having feedback from her fellow experimental colleagues. Currently, she’s coordinating several projects performing functional and pathway analysis at the genomic level. By grouping genes, proteins and other biological molecules into the pathways they are involved in, the complexity of the analyses is significantly reduced, while the explanatory power increases with respect to having a list of differentially expressed genes or proteins.
AlgorithmicsData managementGenomicsImage analysisMachine learningModelingProteomicsSequence analysisStructural bioinformaticsTranscriptomicsDatabaseGenome analysisBiostatisticsScientific computingDatabases and ontologiesApplication of mathematics in sciencesData and text miningGeneticsGraphics and Image ProcessingBiosensors and biomarkersClinical researchCell biology and developmental biologyInteractomicsBioimage analysis
- Exploring pathogenic mechanisms of chronic inflammatory disease: unresolved issues in IL-23/IL-17 biology(YAHIA HANANE - Immunoregulation) - In Progress
- Study of the role of cyclic dimeric guanosine mono-phosphate (c-di-GMP) in the regulation of virulence and biofilm formation in Leptospira interrogans(Gregoire DAVIGNON - Other) - In Progress
- Global BioID-based SARS-CoV-2 proteins proximal interactome unveils novel ties between viral polypeptides and host factors involved in multiple COVID19-associated mechanisms(Yves JACOB - Molecular Genetics of RNA Viruses) - In Progress
After a diploma of statistician engineer from the Ensai (Ecole Nationale de la Statistique et de l’Analyse de l’Information) and a Ph.D in applied mathematics in the Statistics & Genome lab (AgroParisTech), I worked as a developer for the XLSTAT software. I have implemented some statistical methods such as mixture models, log-linear regression, mood test, bayesian hierarchical modeling CBC/HB, … Then I worked as a head teacher in statistics for one year. I was recruited in the Bioinformatic and biostatistic hub of the C3BI (Center of Bioinformatics, Biostatistics and Integrative Biology) in 2014, I am in charge of the statistical analysis and the development of R/R shiny pipelines.
Machine learningStatistical inferenceTargeted metagenomicsBiostatisticsApplication of mathematics in sciencesStatistical experiment design
- Etude de la réponse immunitaire néonatale dans la coqueluche maligne : approche transcriptomique(Soraya MATCZAK - Biodiversity and Epidemiology of Bacterial Pathogens) - Pending
- FLAVIMMUNITY(Giovanna BARBA SPAETH - Structural Virology) - Pending
- Etude de l’évolution des Troubles Olfactifs chez les patients ayant une perte de l’odorat persistante des suites de la COVID-19(Erwan POIVET - Department of Neuroscience) - Pending
Related projects (5)
Chronic inflammatory diseases such as rheumatoid arthritis, inflammatory bowel disease, spondyloarthritis (SpA) and psoriasis cause significant morbidity and are a substantial burden for the affected individuals and the society. An important obstacle to early diagnosis and the development of more specific and effective therapies is the very limited understanding of the pathogenesis of these diseases. In the past years genome-wide association studies have identified many genes that were not known to be involved in pathogenesis, and have linked several genes in immune pathways to inflammatory diseases, indicating that the immune system plays an essential role in the pathogenesis of these diseases. The current challenge is to correlate these genetic variants with the effector mechanisms implicated in pathogenesis, to allow translation of the genetic data into relevant diagnostics and innovative treatment strategies. To meet this challenge, we have designed a clinical study that examines the immune signaling pathways, the transcriptional networks and the genotype in the same SpA patient, in order to establish a link between genetic variation, cellular phenotype and function, and pathology. This approach will advance our understanding of the pathogenic mechanisms, and identify novel and relevant diagnostic tools, therapeutic targets and biomarkers. The long-term outcome of this strategy will be the rational design of specific therapies tailored to the genotype of the patient.
Identification of immune response signatures that correlate with therapeutic responses to TNF inhibitors using machine-learning algorithms
Anti-TNF therapy has revolutionized treatment of many chronic inflammatory diseases, including rheumatoid arthritis, Crohn’s disease and spondyloarthritis (SpA). However, clinical efficacy of TNF-inhibitors (TNFi) is limited by a high rate of non-responsiveness (30-40%) both in SpA and other diseases, exposing a substantial fraction of patients to important side-effects without any clinical benefit. Despite the extensive use of TNFi since many years, it is still not possible to determine which patients will respond to TNFi before treatment initiation. In this study, we have tested the hypothesis that the functional analysis of immune responses may not only improve our understanding of the molecular mechanisms of TNF-blocker activity, but also identify correlates of therapeutic responses in SpA patients. To facilitate the potential translation of our findings into a clinical setting, we have used standardized whole-blood stimulation assays (“TruCulture” assays, Duffy et al., Immunity 2014), and have minimized sources of pre-analytical variability, implementing a highly sensitive and robust pipeline to assess immune functions in patients. To investigate the concept that the immune status of a patient will define their response to TNFi treatment, we have used machine-learning algorithms to identify, in whole-blood stimulation assays, immunological transcripts that correlate with clinical efficacy of TNFi. Our results obtained with a cohort of 67 SpA patients demonstrate that high expression, before treatment initiation, of molecules associated with leukocyte invasion/migration and inflammatory processes predisposes to favorable outcome of anti-TNF therapy, while high-level expression of cytotoxic molecules was associated with poor therapeutic responses to TNF-blockers. These findings may suggest that SpA patients whose immune response is characterized by strong, NF-kB-mediated inflammation are more likely to benefit from TNFi treatment than patients with an active T/NK-cell component. Unfortunately our manuscript describing these results has been rejected by Nature Medicine. However, in her letter the editor mentioned, “Should future experimental data allow you to demonstrate that the identified gene signatures predict response to treatment and outperform previously reported approaches in an independent cohort, we would be happy to look at a new submission…”. We have recruited additional SpA patients over the summer and we are currently in the process of performing the gene expression analysis. The goal of this bioinformatic analysis will be to identify transcripts in stimulated immune cells that predict therapeutic outcome in a training set of patients using machine-learning algorithms and validate the findings in a replication cohort.
Mitochondria are double-membrane bound organelles that are essential in every tissue of the body. They are metabolic hubs and signalling platforms that are deeply integrated into cellular homeostasis. The functions of mitochondria are intimately linked to their form, which is regulated by a balance of membrane fusion and fission: dynamin-like GTPases OPA1 and MFN1/2 perform membrane fusion and DRP1 regulates membrane fission. Mutations in mitochondrial genes cause a pleiotropic spectrum of clinical disorders whose underlying genetic, morphological and biochemical defects can be easily studied in skin fibroblasts generated from patient biopsies. The morphology of mitochondria is inextricably linked to its many essential functions in the cell and we are interested in understanding the relationship between mitochondrial shape changes and metabolism in the context of acquired and inborn human diseases. Balanced fusion and fission events shape mitochondria to meet metabolic demands and to ensure removal of damaged organelles. Mitochondrial fragmentation occurs in response to nutrient excess and cellular dysfunction and has been observed in mitochondrial genetic diseases and is thought to play an important role in the development of disease. The physiological relevance of mitochondrial morphology and the mechanisms that regulate mitochondrial dynamics are incomplete and so we have set out to find ways to rebalance mitochondrial dynamics in genetic diseases. We recently developed imaging and informatics pipelines to allow for the automated, rapid, reliable quantification of mitochondrial morphology in human fibroblasts. We applied this new technology in the context of genome-wide siRNA screens in immortalized fibroblasts from an OPA1 patient with dominant optic atrophy and control fibroblasts to identify candidate genes able to reverse the mitochondrial fragmentation phenotype associated with mitochondrial dysfunction in patient cells. We have performed the same genome-wide screen in healthy immortalized control fibroblasts. Together, these studies will help us identify lists of genetic modifiers and therapeutic targets that can be investigated further using cell biology and biochemical tools in the lab.
Assessing the role of gut microbiota in spondyloarthritis patients and impact of anti-TNF treament on its composition
Our hypothesis is that gut microbiota could define predictive markers of response and tolerance to biologics. A. Preliminary results: gut bacteria predicting response to TNF blockers. A proof-of-concept study has been performed on 58 patients who were recruited according to the following criteria: active disease despite NSAIDs intake; no history of inflammatory bowel disease; no antibiotics intake within 3 months prior recruitment. Bacterial 16S rRNA gene sequencing region was performed on stools samples before and after TNF-blocker treatment. Diversity metrics and custom LefSe were used to explore the relationship between the composition of the intestinal microbiota and the efficacy of TNF-blockers. A lower alpha diversity at baseline was unexpectedly associated with better treatment response, HLA-B27 genotype and smoking behavior. Meanwhile, beta diversity was associated with smoking behavior and HLA-B27 genotype before and after treatment. Beta diversity at baseline was associated with the BASDAI index after treatment, and the response to the treatment. These results indicate a potential regulatory role for the gut microbiota on the underlying mechanisms involved in the response to TNF-blockers. Moreover, a LefSe-like approach identified 6 bacterial species as potential biomarkers for the treatment response, despite the absence of global changes (beta diversity) in the microbiota composition following a 3-month TNF-blockers intake. B. Current project: ITS2 fungal rDNA sequencing analyses In order to establish a causal link between host-microbe interactions and clinical efficacy of anti-TNF, we expect the following research endpoints from our experimental and translation approach (cohorts/clinical trials): 1. Defining the impact of anti-TNF on fungal microbiota and on the relative representation of fungal/bacterial components 2. Defining correlations between gut fungal composition and clinical outcome, with the aim of identifying stool microbial fingerprint of durable responses and/or primary resistance to anti-TNFα in SpA patients. The analyses will be performed on the same 58 SpA patients that were previously analyzed for their 16S bacterial component. These patients perfectly described regarding their disease characteristics, demographics, ongoing treatments, response to anti-TNF after a 3-month treatment period.
HPV RNA-Seq, a novel and unique molecular test that allows detection of pre-cancerous lesions of the cervix from cervical samples.