Using machine learning to explain the fitness landscape of an enzyme involved in antibiotic resistance

An organism’s genome contains the information necessary for its development. Part of this information is used as “instructions” for
cells to synthesize molecules called proteins, which perform most of the cell’s functions. However, over time, changes to the
genomic information can occur. These changes (mutations) can have all kinds of effects: some might cause diseases, others
confer antibiotic resistance to bacteria. Thus, much work has focused on studying what determines the effects of mutations. We
selected a protein involved in antibiotic resistance and are experimentally testing the effects of all possible mutations on its
sequence to understand the determinants of their effects. We believe these determinants are a combination of several factors:
how much of the protein the cell produces, the type of mutation, etc. Using machine learning and integrating data from several
studies, we aim to identify the underlying patterns and relative importances of each of these factors.

Faculty Supervisor:

Christian Landry

Student:

Partner:

Weizmann Institute of Science

Discipline:

Life Sciences

Sector:

Life Sciences (not health); Health and Related Sciences & Technology; Artificial Intelligence; Quantum Science

University:

Université Laval

Program: