It has been well documented that expenditures in technologies related to health informatics and biomedical research have increased dramatically over the past decade. Alberta is an established leader in this regard, investing significantly in the many aspects of ‘eHealth’. This is reflective of Alberta’s priority to continue to improve the quality and sustainability the healthcare system and contributing research in the province. As such IBM Alberta CAS is focused on exploring the many opportunities that exist to extract knowledge from the many sources of data through technologies such as machine learning, data streaming, image processing and the like.

Our portfolio of current and emerging projects in Health and Bio-systems includes:

Accelerated Computational Drug Design for Oncology and Virology

The broad objective of this project, in collaboration with the University of Alberta, is the discovery of new therapeutic agents in the fight against cancer and infectious diseases such as hepatitis.  The goal is to attempt to solve the problem of developing more specific, more efficacious and less toxic chemotherapeutic agents targeting disease for which bio-molecular targets are known through a combination of rational drug design and virtual screening, eventually followed by laboratory validation. Our component of the project will focus on the first two stages using representative test cases of the process, models and techniques.

Collaborators:
Dr. Jack Tuszynski, Professor (Department of Physics, University of Alberta, Allard Chair, Division of Experimental Oncology at the Cross Cancer Institute)
Data Centric Systems, IBM Thomas J. Watson Research Center

Predictive Modeling of Mental Disorders Using fMRI Brain Data

Discriminating between mentally disordered and normal individuals can be a trivial task, but predicting the risk of developing a psychiatric disorder, especially at the early stages of disease onset, is a far more difficult task. In the recent years, the availability of large datasets of brain functional magnetic resonance images (fMRI) under healthy and clinical conditions has provided the opportunity to study brain diseases and to elucidate the underlying neural origins. With the ever increasing power of computers, machine learning approaches have gained growing success in extracting information from these large datasets to identify patterns in healthy versus unhealthy brains. The aim of this project is to use machine learning methods to find reliable patterns in fMRI brain datasets of mental disorders that can differentiate between healthy and patient subjects. Particularly, the team is focused on using probabilistic graphical models (PGMs) that are powerful tools to study brain (dys)functions because of their ability to learn and perform inference in large networks, and using PGMs, build predictive models to determine the odds that a subject has a specific disease, given the brain image.

This project has tremendous commercial potential given the worldwide applicability of the issues tackled and the vast financial investments made in the areas of health informatics and cancer research.

Collaborators:
Dr. Russ Greiner (University of Alberta)
Dr. Matt Brown (University of Alberta)
Dr. Mina Gheiratmand, PDF (University of Alberta)
Dr. Andrew Greenshaw (Department of Psychiatry, University of Alberta)
Dr. SerdarDursun (Department of Psychiatry, University of Alberta )
Dr. Raj Ramasubbu (Department of Psychiatry, University of Calgary)
IBM Computational Biology Group, IBM Thomas J. Watson Research Center