DR-187 Discovering non-coding driver alterations and the clinical utility of mutational signatures

Cancers develop by the accumulation of “driver” genetic alterations providing tumor cells a growth advantage. Whole genome sequencing allows to identify these driver events, but also the mechanisms at the origin of genome insults. Our project has two major aims. Firstly, we will use deep learning approaches to discover driver alterations in the non-coding part of the genome that remains poorly understood. Secondly, we will analyze mutational signatures to discover genomic instability phenotypes related to treatment response. Leveraging the tremendous resource of HMF genomic and clinical data, we hope to identify new oncogenic mechanisms and therapeutic vulnerabilities.

Eric Letouze Cordeliers Research Center France