Deciphering mechanisms of epigenetic alterations in cancer using 3D-genomics-informed deep learning

Principal Investigator: Mathieu Blanchette
Theme : Artificial Intelligence (AI) / Health
Competition : Omics Data Against Cancer (ODAC) Competition
Status : Completed
Start : Oct. 1, 2020
End: Sept. 30, 2022
Budget : $300,000.00



Many types of cancer, including several types of brain cancer in children, are caused by mutations that modify the way cells are able to access their genome - the so-called epigenetic state of the cell. The impact of these, often incurable, cancers on patients and their families is tremendous. Determining which mutations may have this type of dire consequences is critical to help oncologists develop better diagnostic tests and better treatment regimen. Interestingly, these mutations often impact the way the DNA of the cell is folded within its nucleus, preventing it from properly executing its developmental program. In this project, we will build upon recent advances in machine learning (e.g. graph neural networks) to help identify those mutations and predict their epigenetic consequences. We will further use these approaches to better understand when these mutations occur, and their potential causes.

The tools thus developed (algorithms, software) here will be made available to the biomedical research community who will be able to use them to study the links between genetic mutations and epigenetic alterations and their role in the development of several types of cancer.


Lead Genome Centre: Génome Québec 

Partners: IVADO, Oncopole


Nada Jabado McGill University
Josée Dostie McGill University
William Hamilton McGill University


Michael Taylor Sick Kids Hospital, Toronto
Guillaume Bourque McGill University