I am a research engineer with a background in applied Mathematics. I have been working at Institut Pasteur since 2017, as a member of the Statistics Group of the Hub of Bioinformatics and Biostatistics. I work on various applications of statistics on data generated by Institut Pasteur research units including:
- statistical analysis of high-dimensional data
- differential analysis of transcriptomics and proteomics data
- functional analysis of transcriptomics and proteomics data (GSEA)
- analysis of RNA-Seq bulk and single-cell data
- cellular deconvolution (cell proportion estimation) from bulk RNA-Seq data
- development of Shiny applications
I also work in collaboration with the animal facility to support scientists in experimental design optimization (3R) as well as sample size computation and power analysis. As a member of the Hub, I am involved in the development of a Shiny application named SHADE to assist researchers in designing their experiments.
Before joining Institut Pasteur, I received a PhD in Statistics in 2015 from the Université de Rennes 1, working at Agrocampus-Ouest under the supervision of David Causeur. My PhD was about variable selection for supervised classification in high dimension, when predictors are correlated. I applied the theoretical results of my PhD to ERP (Event Related Potentials) data in Psychology. During my postdoctoral position at INRIA at Grenoble I worked on robust nonlinear regression, using mixtures of regression and heavy tailed distributions, finally applying the results to planetology data.
Links R code to run simulations of the article Devijver E. and Perthame E., Prediction regions through Inverse Regression, Journal of Machine Learning Research 21 (2020) 1-24, available here