Marc-André Legault, a newly appointed professor in the Faculty of Pharmacy at Université de Montréal, wants to improve what’s known about the interplay between genetic variation and drug response.
He is combining genetic statistics, genetic epidemiology and machine learning to identify which patients are most likely to benefit from a given drug and to minimize its side effects.
And he believes his research has the potential to revolutionize personalized medicine.
A passion for research from day one
Legault became hooked on research during his first year of undergraduate studies in biomedical sciences at UdeM.
He interned at the School of Optometry and worked in a neuroscience lab. Those experiences left a lasting impression: he remembers throwing himself wholeheartedly into his first real research project.
“I got to conduct my own experiments, write up the results and interact with the rest of the team,” he recalled.
Then a pivotal encounter with a bioinformatics researcher changed his career path.
“He told me about the emerging field of bioinformatics and how useful it was for solving concrete problems in biomedical research,” said Legault.
Intrigued by the potential of bioinformatics – and motivated by his passion for new technologies – he decided to specialize in the field.
Legault joined Marie-Pierre Dubé’s lab at the Montreal Heart Institute, where he honed his skills in developing bioinformatics. Among other things, he worked on comparing methods for detecting structural genome variation. He also co-authored his first scientific publication, a groundbreaking work in which he used genetic data from monozygotic twins to compare variation-detection algorithms.
Transforming personalized medicine
Legault explores cutting-edge approaches in pharmaco-omics with the goal of transforming personalized medicine. Rather than limiting himself to analyzing a single gene or protein, he takes a more comprehensive approach that considers all the molecules in the body.
“In genetic epidemiology, we focus on simple biomarkers such as cholesterol levels to assess the risk of heart disease,” he explained. “With these new omic approaches, we analyze the effects of drugs on all of the body’s proteins.”
For his doctoral studies, Legault analyzed the effects of 26,616 proteins on 1,746 phenotypes using data on more than 413,000 individuals in the UK Biobank, a database containing detailed genetic, phenotypic and health information on over half a million British citizens.
His results, which are available on the ExPheWAS platform for use by other researchers, are leading to a better understanding of the links between genetic variation and various diseases.
The ultimate goal of Legault’s research is to draw a complete molecular portrait of a patient’s state of health by integrating data on proteins and other biomarkers. This will eventually enable health practitioners to select the most appropriate drug for each individual based on their unique biological profile.
“We will be able to prioritize treatments, based on personalizing drug choices according to the patient’s current condition,” Legault said.
Mimicking a drug’s effect
A key aspect of Legault’s research is modelling the effects of drugs on people who are not taking them and linking those effects to genetic variation.
“If a person has a genetic variant that naturally reduces an enzyme’s activity, this could mimic the effect of a drug designed to inhibit that enzyme,” he explained.
These models are proving useful in drug development, especially for designing clinical trials and predicting the long-term effects of drug use.
At Mila – Quebec Artificial Intelligence Institute, where he is an associate researcher, Legault is developing bioinformatic tools and machine-learning algorithms to analyze the impact of genetic variation on treatment response.
One focus of his research at Mila is drug repositioning – that is, using an existing drug to treat other diseases. For example, a drug developed for Crohn’s disease could be used to treat arthritis, based on analysis of the drug’s effects on omic profiles.
Useful for treating children
Legault is also a researcher at the Azrieli Research Centre of CHU Sainte-Justine, where he focuses on predicting which drugs normally prescribed for adults could also be useful for treating children.
This presents a major challenge for a researcher used to working with supercomputers and massive databases.
“With adults, it’s common to have data on more than half a million, even a million, people,” Legault said. “But it’s extremely rare to have such large cohorts for children. Understanding the effects of a drug at the molecular level in children will help us prioritize treatments tailored to their specific needs.”
In addition to his research, Legault enjoys sharing his knowledge with students at Université de Montréal. This winter, for example, he’ll be teaching Mendelian randomization, a method he uses in his own research.
“Teaching stimulates research, and I hope to pass on my passion for research to my students,” he said.
For Legault, the greatest reward is seeing students inspired by his teaching join his lab, ready to contribute to research and lead their own innovative projects.
