Welcome!














Avatar

Marie Perrot-Dockès

Maitre de conférence (Assistant professor)

Université de Paris

Biography

I’m currently an assistant professor at the MAP5 laboratory and the STID departement of IUT de Paris, which is part of Université de Paris. I’ve done a PhD in AgroParisTech supervised by Céline Lévy-Leduc, Julien Chiquet and Laure Sansonnet. After thtat, I was in a postdoctoral position, between Sorbonne Université and Toulouse institut of mathematics, mentored by Etienne Roquain and Pierre Neuvial.

  • Research: My PhD titled Regularized methods to study multivariate data in high dimensional settings : theory and applications aims at selecting variable of interest in the multivariate linear model when the number of responses is much higher than the sample size. My postdoc aims at finding upper bounds of the number of false discoveries for all the possible sets of selected variables in a multiple testing framework. These upper bounds have to be valid with high probabilities. Most of the existing bounds are too conservative and a crucial challenge is to build new post hoc bounds with better performances.

  • Teaching:

    • During my PhD I have given 192 hours of directed works on statistical inference and linear modeling. The students were the first and second years of the engieneering school AgroParisTech.
    • During my postdoc I’ve given a 30 hours course on probabilities and statistics to the Master students of Polytech school of Paris.
    • Since 2020 I teach in IUT de Paris.
  • Association As a member of the young group of the French statistical society I co-organised a meeting of the young European statistician, a session of the JDS (French statistical days) 2019, and one session and two workshops for the JDS 2020.

Interests

  • Variable selecion
  • Multiple testing
  • Post hoc inference
  • Application in genetic
  • Application in immunology

Education

  • PhD in Statistics, 2016 - 2019

    AgroParisTech / Université Paris-Saclay

  • Statistician Diploma, 2012-2015

    ISUP

  • Classe Préparatoire, 2010-2012

    Lycée Charlemagne

Experience

 
 
 
 
 

Postdoctoral position

LPSM/ IMT

Oct 2019 – Present Paris / Toulouse
Responsibilities include:

  • Courses at PolyTech for the master student
  • Theoretical research on post hoc bounds on the number of False discoveries
  • Collaboration with Pierre Gestraud of the curie institute on CNV dataset
  • Collaboration in Astrophysics.
 
 
 
 
 

PhD student

AgroParisTech

Oct 2016 – Sep 2019 Paris / Grignon
Responsibilities include:

  • Courses at the student of the 1 and 2 years of AgroParisTew<ch
  • Theoretical researche on multivariate models
  • Collaboration with immunologist from the Curie Institut
  • Collaboration in plant molecular physiology.
 
 
 
 
 

Research Ingenieur

Institut Curie

Sep 2015 – Sep 2016 Paris
Statistical analysis to study the dialog between Dendritic Cells and Th Lymphocyte.

Recent & Upcoming Talks

Nouvelles bornes post hoc dans un cadre structuré

Les procédures classique de test multiples utilisent un seuil $\alpha$ définit à l’avance et rejettent un certain nombre d’hypothèse en …

Post hoc false positive control in Hiden Markov Models for spatially structured hypotheses

Introduction to the post hoc false positive control

In this presentation we study general linear model (multivariate linear model) in high dimensional settings. We propose a novel …

Méthodes régularisées pour l’analyse de données multivariées en grande dimension : théorie et applications.

In this presentation we study general linear model (multivariate linear model) in high dimensional settings. We propose a novel …

Recent Publications

Quickly discover relevant content by filtering publications.

A Quantitative Multivariate Model of Human Dendritic Cell-T Helper Cell Communication

In this article we used statistical model to described rules that explain communication between dendritic cells and T helper cells. …

Regularized methods to study multivariate data in high dimensional settings: theory and applications.

Theoretical, numerical and applied studies of a novel variable selection approach in the framework of multivariate linear models taking …

A multivariate Th17 metagene for prognostic stratification in T cell non-inflamed triple negative breast cancer

A diversity of T helper (Th) subsets (Th1, Th2, Th17) has been identified in the human tumor microenvironment. In breast cancer, the …

Estimation of large block covariance matrices: Application to the analysis of gene expression data

Motivated by an application in molecular biology, we propose a novel, efficient and fully data-driven approach for estimating large …

A multivariate variable selection approach for analyzing LC-MS metabolomics data

Omic data are characterized by the presence of strong dependence structures that result either from data acquisition or from some …

Contact