Improving Robust High-Dimensional Causal Inference and Prediction Modelling

> français

Collaborative Research Team Project: 2021-2024

Recent advances in “-omics” technologies allow the simultaneous quantitation of thousands of features, revolutionizing the way that scientists can measure pathogenic processes or responses to therapies. Despite their potential to improve diagnostics and prediction methods in routine clinical care, these high-dimensional multi-faceted data still present significant challenges, including measurement errors, outliers, multivariate responses, and complex correlation structures. For example, genomics data, imaging scans and years of standard clinical lab tests are sometimes available from the same subject. Furthermore, many of these features may be inaccurately measured, as a consequence of technical limitations or errors in recording data. A central goal of this Collaborative Research Team is to develop and establish an advanced analytical framework for the study and integration of complex data in biomedical sciences, including advanced regularized regression methods, robust regularized instrumental variable methods, and matrix-valued causal models, all for high-dimensional settings. These advancements are essential for building useful models in precision medicine.

Team Leaders:

Gabriela Cohen-Freue – University of British Columbia
Celia Greenwood – Lady Davis Institute for Medical Research, McGill University


Sahir Bhatnagar – McGill University
Tom Blydt-Hansen – University of British Columbia
Dehan Kong – University of Toronto
Karim Oualkacha – Université du Québec à Montréal
Brent Richards – McGill University 
David Soave – Laurier University 
Linbo Wang – University of Toronto
Zhaolei Zhang – University of Toronto

Comments are closed.