Deep learning for precision medicine by joint analysis of gene expression profiles measured through RNA-Seq and microarrays

CO-FUNDED PROJECT BY IVADO 

Immunotherapy is a recent and effective treatment for certain types of cancer. Molecules called major histocompatibility complexes class I (MHC I) play a central role in this type of therapy. Various MHC I are found on the surface of cells and serve to present the immune system with a range of short peptide fragments denoting the internal activity of the cell. These peptides are called MHC I-associated peptides or MAPs.

Immune system cells are able to determine if MPAs are “harmless” or “abnormal” and then eliminate those found to be abnormal. This is how the immune system monitors and controls the presence of infected cells or the spontaneous appearance of cancer cells. Immunotherapy takes advantage of this system. One strategy is to develop vaccines using MAPs that are specific to the targeted cancer. These vaccines stimulate the patient’s immune system to eliminate cancer cells. The identification of MAPs specific to a patient’s tumour is a key step in the development of effective and personalized cancer immunotherapy. This project aims to develop a new computational approach to identify MAPs using deep neural network methods that will be applied to large RNA sequencing and tandem mass spectrometry (MS/MS) datasets. This project is designed to make it easier to identify MAPs and enhance the reliability of these identifications, in turn increasing the number of potential MAPs to be targeted for cancer vaccine development.

Lead Genome Centre: Génome Québec

Partner: IVADO

Co-investigators: