Developing a tool to identify epitopes for drug therapy using semi-supervised and supervised machine learning

The COVID-19 pandemic has been caused by a novel coronavirus, SARS-CoV-2. Viral infections cause the human immune system to respond in two ways: a fast, non-specific response, and a slow response that produces antibodies that will target the infecting virus. Each antibody will recognize a small segment of the virus, called an epitope. If the best segments to target can be identified, it is possible to manufacture antibodies to test for the infection and help the immune system to boost its response when fighting the infection. Identifying these epitopes manually is a labor-intensive process, that is time consuming and expensive. In the current project, we will use machine learning techniques to gather data of known epitopes and predict new epitopes to new infections. We expect this work will revolutionize the discovery of antibodies when fighting new infections in an epidemic or pandemic context.

Faculty Supervisor:

Eldad Haber

Student:

Philipp Witte

Partner:

Genomica

Discipline:

Geography / Geology / Earth science

Sector:

Professional, scientific and technical services

University:

University of British Columbia