Mots-Clés

Data integration Multi-omics Missing data Deconvolution Clustering

Description

Personalized medicine is an emerging approach in healthcare that gained momentum with the advent of genomic technologies in the early 2000s. It focuses on tailoring medical decisions to an individual patient’s predicted response. A notable example is the classification of breast cancer into five subtypes based on gene expression profiles, as well as the development of monoclonal antibodies, which are particularly effective in HER2-positive patients [Tam16].
For personalized and precision oncology to reach its full potential, we must deepen our understanding of tumor evolution at the patient level. This involves identifying the initial events that trigger cancer formation and the successive driver events that lead to therapeutic resistance, relapses, and uncontrolled tumor growth.
As sequencing technologies become more accessible, it is now common practice to use gene panels sequencing at the time of diagnosis — that is sequencing only a selection of genes, known to be associated with cancer through mutations, expression changes, or copy-number alterations. This helps to identify potential abnormalities in tumor cells that can guide treatment and prognosis. Translational research labs are actively developing new sequencing panels to uncover both known and novel events at the RNA and DNA levels, with the goal of improving patient outcomes. This process generates large volumes of heterogeneous data, collected at multiple time points from different patients.
The general aim of the postdoc project will be to develop a statistical approach to identify the clonal composition of the tumor from panel-sequencing data. To that effect, the project will be structured in three main aims, detailed in the url link.