Comparison of algorithms for predicting drug-target interactions and performance benefits of ensemble learning

The goal of this project is to compare the results of several state-of-the-art artificial intelligence (AI) models in the field of drug discovery. Specifically, the intern will recreate several models whose creators claimed are successful at predicting which drugs would interact with what proteins in the human body. In addition, the intern will design a model of their own based on existing successful models and their own ideas. Lastly the output of these various models will be combined to give a single, accurate prediction regarding which pairs of drugs and proteins will interact. This project will help reproduce and validate the results of these models, provide a novel solution to the problem they are designed to solve, and help evaluate the benefits of combining the results of various models in this fashion. In addition, this work will form a valuable addition to Modelis’ computation drug discovery pipeline.

Faculty Supervisor:

Ioannis Mitliagkas

Student:

Partner:

Modelis inc.

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

Université de Montréal

Program: