When Michel Dumontier and James Green speak about their latest project, their faces light up. If successful, the project could lead to groundbreaking developments in biomedicine and help develop powerful new diagnostic tools.
It all started when Dumontier, currently an assistant professor in bioinformatics at Carleton University, was working on his post-doctoral research in Toronto three years ago. His research focused on proteomics (i.e., the study of proteins) and, specifically, the use of mass spectrometry, an analytic method that can identify proteins from mixtures.
Mass spectrometry has become important in the discovery of biomarkers—proteins that appear only in the context of disease—and can therefore provide valuable information in the development of new therapeutic drug treatments.
Currently, mass spectrometry devices rapidly collect raw data and the slow process of protein identification is made afterwards. This de-coupling of the collection and identification is problematic because an accurate identification might require additional information to be collected, which cannot be done because the sample is completely used up.
“If the collection and identification were done simultaneously, then rarely seen proteins that have a role in human health and disease could be accurately identified,” explains Dumontier.
In the last year, he and Green, an assistant professor in the department of systems and computer engineering, have been working closely to identify computer technology that would allow mass spectrometers function to simultaneously collect and analyze data.
In particular, they have been investigating novel computer architecture called Cell BE, which was designed by Sony, Toshiba and IBM and can be found in every Playstation 3. Dumontier and Green say their long-term goal is to incorporate Cell BE technology into clinical applications and to use their findings for diagnostic purposes such determining whether a person has a disease or needs a biopsy.
“Although there are many effective diagnostics for certain diseases, we want to identify those that aren’t available yet and design new targeted clinical applications,” says Dumontier.
The Cell BE is a single chip with nine processors. It is unique because it has one power processing element (PPE) and eight synergistic processing elements (SPEs).
“The PPE acts like a manager and hands out jobs to the SPEs, keeping them busy and sending them data,” explains Green. “The SPEs are highly specialized to be effective at streaming data in and out, and doing mathematical operations. This chip gives a lot of opportunities for achieving massive processing speed-up, but it’s a real challenge to program.”
He adds that both he and Dumontier hope Cell BE will speed up protein-related experiments and also enable new protein identification algorithms.
So far they have been using a virtual loaner program from IBM to conduct their experiments on a server in Texas, and will be attending an advanced workshop on the technology to be held at Carleton University in May. But—thanks to recent funding—they expect to be working with real hardware in the near future. They have received $114,628 from the Canada Foundation for Innovation and are waiting to receive a matched amount from the Ontario Research Fund.
To create a device that will carry out existing proteomics experiments faster and better, and enable new forms of proteomic analysis and clinical diagnostics.
Initially, researchers in biochemistry and molecular biology laboratories. Later, patients seeking medical diagnostics.
Through the use of Cell BE, the researchers aim to accelerate protein data analysis to the point where it will be fast enough to do simultaneous data collection and analysis. This would lead to new, better and more accurate results.
To develop innovative medical diagnostic tools.