Société Française de Bioinformatique

Mots-Clés

single cell protein expression machine learning deep learning flow cytometry imaging flow cytometry supervised learning unsupervised learning

Description

The Centre de Recherche sur l’Inflammation (CRI) is France’s biggest research center in pathophysiology. The CRI brings together clinicians, biologists and data analysts to understand chronic inflammatory diseases and cancer. This postdoctoral position is open in the Paradis-Sinkus team “From Micro to Macro in cancer development (MicMac)” whose focus is to understand the heterogeneity of chronic inflammatory liver diseases across patients, and to correlate it with the emergence and biology of cancer. The succesful candidate will be supervised by Dr. Etienne Becht, a computational scientist who recently joined the group and was since awarded an ANR - Young Researcher grant.

Flow cytometry is a well established technique to study protein expression at the resolution of single cells, but is limited in the number of assayed proteins. The project builds upon previous work from the group that demonstrated how non-linear regression machine learning models could fruitfully be combined with parallel flow cytometry to drastically increase the number of proteins assayed (from about 20 to more than 300).

The candidate will contribute to the following axes :
–Extending our previous work from conventional to imaging flow cytometry (ImageStream Mk2 cytometer) using suitable neural network architectures
–Applying supervised learning methods to phenotype (classify) cells in the context of flow cytometry
–Develop and assess feature selection methods for flow cytometry
–Analysis of “cellular atlases” from high-throughput surface protein expression data from human tissue samples