Description
Date depot: 7 septembre 2023
Titre: Deep learning approaches as predictors of the cell’s regulatory networks
Directrice de thèse:
Laura CANTINI (Machine Learning for Integrative Genomics)
Directeur de thèse:
Gabriel PEYRÉ (DMA/ENS)
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Intelligence artificielle
Resumé: Single-cell transcriptomics (scRNA-seq) has revolutionized biology and medicine by unrevealing the diversity of the cells constituting human tissues. However, the complex task of inferring gene regulatory networks has not yet profited from this revolution and tools have not been predictive enough to reach any usability. Knowing such a network for each cell type is expected to provide a more comprehensive view of the cell’s states and explain some of the different cellular phenotypes and responses of the cells.The aim of this Ph.D. project is to develop novel approaches using deep learning and specifically novel graph neural network approaches on large scRNAseq datasets, to assess their predictability in high-quality benchmarks and package them as an open-source Python library.The methods developed during this project will be applied in collaboration with wet-lab biologists, in order to derive new biological knowledge from their in-house data.This PhD project will impact both computational fields and biomedical fields, by developing rigorous methods that can reliably maximize the information extracted from complex multimodal datasets. In particular, applications of the tool will contribute to personalized medicine.
Doctorant.e: Kalfon Jérémie