Collaboration Hub

WHO: Dr. Ajay Shah with Sunnybrook Health Sciences Centre 

PROJECT: "Predicting conversion to total knee arthroplasty following anterior cruciate ligament reconstruction: a machine learning approach." We have a large dataset including MRI results (tabular form), lab tests, clinical exam findings, patient-reported outcome measures, on a cohort of patients who had ACL reconstruction in 2005-2007. We would like to employ ML techniques to create a predictive model to determine:
1) preoperative features of highest importance to predict target (total knee replacement at followup)
2) create 'calculator' for assessing risk of arthritis and knee replacement

CURRENT PROJECT COLLABORATORS: Dr. David Wasserstein, Dr. Paul Marks, (Sunnybrook Orthopaedic Surgery faculty)

LOOKING FOR COLLABORATORS WITH:
• ML skills (NN, pytorch, elasticNet, RF)
• Creativity - suggesting ideas/refinement of models
• Honesty - when models aren't well-suited, etc.

NOTES: Involvement in this process will lead to authorship on the publication of this manuscript. The estimated timeline is 4-6 months from analysis to submission.

CONTACT:
Dr. Ajay Shah
drajay.shah@mail.utoronto.ca
Tel: 647-782-5990

Posted March 20, 2025

Postdoctoral Scholar – Kidney Health Informatics & Data Science

Job posting coming soon 
Location: Calgary, Alberta, Canada
Duration: Two years (with possibility of extension)

The Department of Medicine at the University of Calgary and the Alberta Kidney Disease Network (AKDN) are seeking a Postdoctoral Scholar in Kidney Health Informatics and Data Science. This position offers a unique opportunity to contribute to cutting-edge research in predictive analytics and risk assessment for people with kidney disease.

Key Responsibilities:

• Design of advanced predictive models (machine learning, statistical learning)
• Creation and evaluation of risk prediction tools for kidney disease using large-scale administrative, clinical, and laboratory databases

Qualifications & Requirements:

• PhD (completed within the last three years) in research methodology, statistics, or data science, with a background in mathematics or computer science
• Strong analytical skills and experience with machine learning, predictive modeling, and large datasets
• Competitive publication record and academic achievements
• Willingness to apply for external fellowship funding
• Fluent in English and eligible to reside in Canada

Why Join Us?

• Conduct high-impact research under the mentorship of Dr. Ping Liu and Dr. Pietro Ravani
• Work with a leading research network (AKDN) and contribute to high-impact publications
• Receive mentorship, fellowship funding and professional development support
• Engage in a collegial and supportive research environment

Application Process:

To apply, email the following to Dr. Ping Liu (ping.liu1@ucalgary.ca):

• Letter of interest
• Curriculum vitae (CV)
• Relevant publications
• Contact information for three references

Researchers: Dr. Sabri Soussi and Dr. Claudia dos Santos
Institution: University of Toronto
Project title: AI in Shock and acute Conditions OutcOmes Platform (Shock CO-OP)

Notes: If interested, please email ASAP. Preferred applicants are those who apply before May 12, 2024, but we will look at applications after that as well. 
sabri.soussi@uhn.ca;  claudia.dossantos@unityhealth.to

Project description: This collaborative research initiative between Université Paris Cité (France) and the University of Toronto represents a groundbreaking effort to redefine the classification of circulatory shock through the application of artificial intelligence (AI) approaches for the integration of high dimensional biomarker data.

The collaborative project, known as the AI in Shock and acute Conditions OutcOmes Platform (Shock CO-OP), seeks to enable a paradigm shift in the understanding of circulatory shock clinical syndrome by identifying distinct subclasses based on physiological/molecular profiles (i.e., subphenotypes).

Our international/multidisciplinary research group already identified distinct subphenotypes in septic shock patients with different outcomes and inflammatory and cardiovascular patterns. By leveraging existing clinical and biomarker data from complementary cohorts/clinical trials with biobanks in Europe and North America, our research group aims to further uncover novel insights into the pathophysiology of circulatory shock independently from its etiology (e.g., infection, myocardial infarction, trauma, major surgery) by identifying distinct biomarker-driven subphenotypes. This will allow the development of biomarker/subphenotype-based therapies to improve short-/long-term patient-centered outcomes (i.e., precision critical care medicine).

Desired skills in collaborators: 
• Seeking Graduate Students for 1-2 Years
• Good programming skills in Python and/or R especially for:
1) high dimensional metagenomic, transcriptomic and proteomic data management and analysis.
2) Longitudinal and functional data analysis and unsupervised machine learning and model-based clustering (i.e., mixture modeling) for subphenotyping purposes in heterogeneous populations.
-Proactive, team player, self-motivated and able to learn new approaches 
-Master and PhD candidates

The Imagine Institute and Necker Hospital have developed a data warehouse of 1 million pediatric patients with “all for all“ phenotyping based on their raw electronic health record data. Histories of rare disease patients (as well as common disease patients) are extracted from patient electronic health records, and clinical images and photographs (e.g., for facial dysmorphia) are are also available. Clinical data can be linked with the -omic data generated at the Imagine Institute. Omic data are from different types: genomic, exome, transcriptome, single cell, and organ-specific such as those derived from Urine-Derived Renal Epithelial Cells (URECs). In addition, collaborators will have the opportunity to work with the Paris AI Institute (PR[AI]RIE). All interested collaborators are requested to contact

Contact: Dr. Anita Burgun

Researcher: Yalini Senathirajah
Institution: University of Toronto
Project description: Interested in collaborations with those with a mature AI recommendation algorithm that could be integrated into the EHR (our research is on user-controlled platforms for EHR redesign)
Project aims: Composable approaches to EHR design may have benefits for human-computer interaction, cognitive support, efficiency, workflow, and other benefits. It also may make it easier to incorporate AI-based recommendations at the point-of-care. We are interested in possibly piloting this approach with those with appropriate tested algorithms seeking to incorporate them into the EHR.
Desired skills in collaborators: Those who have developed AI-based algorithms aimed at clinical care, at an appropriate stage of development/testing.

Contact: Adam Kepecs (Tel: 647-812-6678)
Organization: IntelAGENT
Project aims: This project aims to use Machine Learning to parse a clinician's unstructured free-text patient encounter note in their Electronic Medical Record (EMR) system to identify billable OHIP services. This will alleviate administrative overhead and help physicians spend more time in clinical practice.
Desired skills: Expertise in Natural Language Processing models for unstructured clinical notes.
Other collaborators in this project: Prof. Frank Rudzicz

Researcher: Karim Mekhail seeks skills from another member
Institution: University of Toronto
Website: https://www.mekhaillab.com
Project aims: DNA mobility can contribute to genome stability. Although initially believed to be random, this mobility can exhibit non-random motions consistent with gaining access to repair-conducive factors. The current goal is to predict the motion of DNA and the outcome of repair on a single-cell level.
Desired skills in collaborators: Advanced knowledge of time series prediction models.