Réglage automatique et visualisation interactive d’expériences unicellulaires

Details

Supervisors

Dupont, Pierre

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] en sciences informatiques, à finalité spécialisée

Abstract

enWith the help of algorithms, it is now possible to analyze cellular genetic expression data. Among these algorithms one is called MicroCellClust. It seeks to extract small groups of cells from datasets that sometimes contain hundreds of cells. It does this on the basis of the gene expression data of these cells. This algorithm, although efficient, is difficult to understand for researchers in the field of health. Indeed, it solves an optimization problem under constraints and the researchers do not have the necessary training to fully master this problem. Thus, it is difficult for them to define the parameters which will lead to a solution which will be able to suit them. In the course of this work, two new things were learned at MicroCellClust. To begin with, the role of the meta-parameters of the algorithm was understood. After that, the automatic tuning of these was implemented and integrated into the but to make MicroCellClust accessible to medical researchers. Finally, an interactive platform makes it possible to visualize the results obtained by the execution of the algorithm.