Graph Neural Networks for Drug Discovery at Simmunome Inc.

This project will help Simmunome Inc. better evaluate and explain its machine learning models for biomedical discovery. Simmunome uses biological network data, such as relationships between proteins, genes, pathways, and disease targets, to support decisions about which biological targets may be worth investigating further. The intern will build a structured testing process to compare Simmunome’s current graph-based model with simpler models, standard graph models, newer state-of-the-art methods, and control models where the graph structure is changed or removed. This will help determine when graph-based modelling truly improves predictions. The intern will also develop methods to explain model predictions by identifying the proteins, connections, pathways, or interaction patterns that most influence a result. The expected benefit is a clearer and more reliable evaluation toolkit that helps Simmunome improve its models, understand when predictions can be trusted, and communicate results more clearly to scientific users, clients, or collaborators.

Faculty Supervisor:

Reihaneh Rabbany

Student:

Partner:

Simmunome Inc.

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

McGill University

Program: