Spring 2022: Kevin Dick and Yasmina Souley Dosso have defended their PhD theses! Eric Arezza has defended his MASc thesis! Our lab will be publishing papers at IEEE I2MTC and IEEE MeMeA this spring. Kevin Dick's PhD thesis has been nominated for a Senate Medal. James Green presented a 40-minute workshop on "Machine Learning (ML) Applications in Critical Infrastructure Monitoring" as part of the CEATI Virtual Black Sky Hazards Workshop. Congratulations to incoming MASc student, Anthony Fuller, who has received a Vector AI Scholarship to support his graduate research in Vision Transformers for Remote Sensing.
Fall 2021: PhD student Yasmina Souley Dosso has had a paper accepted to Elsevier Computers in Biology and Medicine on the topic of "RGB-D Scene Analysis in the NICU". PhD student Kevin Dick has published in the Journal of Proteomics with collaborators at Agriculture Canada. Kevin and Daniel had a paper accepted at SPLASH-E describing the integration of Machine Learning research into an undergraduate course. Our lab will also be getting together for a hike in Gatineau Park in October as we transition to post-COVID19 life.
Summer 2021: PhD student Kevin Dick has won the Oustanding TA Award for SYSC4906 - Intro to Machine Learning. Apparently 274 TAs were nominated and only 5 awards were given, so congratulations Kevin! Congrats also to multiple lab members who have had conference papers accepted this summer (Kevin/Josh/Francois CRV2021, Yasmina EMBC2021, Mohsen/Daniel/Yasmina/Josh SAS2021.
Winter 2021: Congratulations to MASc student, Daniel Kyrollos, for winning 3rd place in the Data Day 7.1 poster competition for his poster on A Meta-Model in NLP for Hatefulness. Daniel will also receive an Ontario Graduate Scholarship next year. Kevin Dick was the lead author of a PeerJ paper on "Multi-schema computational prediction of the comprehensive SARS-CoV-2 vs. human interactome". Postdoc Dr. Roy Wang was lead author on a review paper in Metabolites on "Automatic 1D 1H NMR Metabolite Quantification for Bioreactor Monitoring". Kevin Dick and Daniel Kyrollos both published papers in IEEE I2MTC 2021. Kevin Dick participated on an expert panel at Data Day 7.1 on the topic of "Smart Everything". Fadwa Darwaish and Roger Selzler both had papers accepted to IEEE MeMeA 2021.
Fall 2020: Our NSERC Alliance-funded COVID19 research has been higlighted in a recent story.
CUBIC: Carleton University Biomedical Informatics Co-laboratory
At the Carleton University Biomedical Informatics Co-laboratory (CUBIC), we apply machine learning and data science to solve problems in biomedical informatics. We are particularly interested in predicting rare events, or conducting machine learning in the presence of class imbalance. Current projects requiring additional students include an exploration of the use of RGB-D video and pressure-sensitive mats for real-time patient monitoring in the NICU at CHEO, characterizing and mitgating vibrations experienced by neonatal patients during emergency ground and air transport to the NICU at CHEO, and development of novel machine learning methods for analyzing protein structure, function, interaction, and chemical modification. Interested students should have strong software and communication skills proven through academic performance and/or industry experience. Hands-on experience with software development, mahcine/deep learning, computer vision, web development, and statistics is highly valued.
Non-contact Neonatal Patient Monitoring We are working with CHEO to investigate novel patient monitoring technologies in the NICU. We are examining the use of pressure-sensitive mats (PSM) to provide continuous, unobtrusive, and non-contact monitoring of critically ill babies in the neonatal intensive care unit (NICU) at CHEO. We are working with IBM's Centre for Advanced Studies (IBM-CAS), Dr. JoAnn Harrold (Neonatologist @ CHEO), and Mr. Kim Greenwood (Director of Clinical Engineering @ CHEO). In addition to PSM, we are also using multispectral cameras (colour, near-infrared, and depth) to record video of patients from above. From the PSM and video data, we are developing deep learning computer vision approaches for physiological monitoring (HR, RR, etc), characterizing patient movement, and detecting clinical interventions with an eye on semi-automated charting. We have developed a tablet app, such that bedside researchers can annotate all events of interest for up to 6 hours per patient. These gold-standard annotations will be used to develop and validate machine learning approaches to semi-automated non-contact patient monitoring in the NICU.
Estimation of Patient Stress during VR-Therapy Together with Prof. Adrian Chan and clinicians from the Canadian Forces and the Ottawa Hospital Rehabilitation Centre, we are developing novel methods for estimating patients' stress levels (sympathetic activation of the autonomic nervous system) during virtual reality therapy for PTSD, mild traumatic brain injury, and complex pain. Such monitoring will enable clinicians to tailor the intensity of therapy to induce brain plasticity and healing, while avoiding over-stimulating the recovering brain. We are using both gait information, from a VICON motion-capture system, and physiologic signals from a wearable sensor. Deep learning models will be used to analyze these data to generate estimators of stress in real-time, thereby providing an important tool to rehabilitation clinicians.
Quantifying Patient Vibrations During Emergency Transport Each year, thousands of newborns in Canada are transported by air or ground ambulance to receive specialized medical care. These infants are often premature and especially vulnerable during transport. Vibration exposure during transport may contribute risk of serious long-term consequences including brain injury. To decease risk during transport, specialized equipment is used. However, vibration levels experienced by infants and contributing factors are not well understood. This lack of knowledge affects the ability to transport newborn patients in the safest manner possible. With Rob Langlois (MAE), Adrian Chan (SCE), and Stephanie Redpath (CHEO), we are seeking to increase the safety of infant transport by reducing vibration exposure. We propose a research program which seeks to understand how vibrations caused by the road and air environments are transferred to the infant. This will allow us to propose novel methods to reduce vibration exposure and improve the equipment for transporting fragile infants. We will measure vibrations in many transport scenarios including within hospitals as well as in ground ambulances, helicopters, and airplanes. The results will be used to develop test equipment and procedures for evaluating the transport equipment. This research will enable better understanding of the problem and provide a reliable way to test the transport equipment and procedures. It will lead to new tools for planning routes that minimize vibration and for monitoring patient vibration.
Semi-Supervised and Species-Specific Prediction of microRNA. microRNA are short RNA molecules that play an important role in post-transcriptional gene regulation. Our collaborators are continuously sequencing new species and wish to identify novel microRNA within these new genomes. However, most widely-used microRNA prediction tools are only effective on human data. We have developed SMIRP, a framework for the creation of species-specific predictors of microRNA from genomic sequence. We have achieved up to 500% increases in sensitivity at precisions of up to 90% when compared with existing methods. SMIRP has been applied to study numerous genomes including turtles, slime moulds, and a snail. We are now developing methods to leverage transcriptomic RNA-Seq data as this continues to becomes more accessible to experimental researchers.
Post-translational modification. While progress continues to be made on the prediction of structure from sequence, knowledge of a protein's structure may not be sufficient to discern its function. For example, most proteins undergo some form of post-translational modification (PTM) following initial synthesis which may have a profound impact on protein function. Our lab is therefore working to develop intelligent predictors of important PTM's such as sumoylation and phosphorylation. Iterative prediction of protein function and structure is a long term goal as well.
Check out our YouTube channel for recent conference presentations by members of the cuBIC lab.
Bioinformatics web services. Please click here for a partial list of web services developed by our lab.
Areas of Research Interest
My research focus has been in the following areas:
Real-time patient monitoring, using pressure-sensitive mats, multi-modal video, and other sensors.
Machine learning, pattern classification, data mining
Bioinformatics, proteomics, and prediction of protein structure, function, interaction, and post-translational modification
Development of novel assistive technology and devices
Current projects include:
Neonatal patient monitoring using pressure sensitive mats (collaboration with IBM-CAS)
Classification of audiograms for tele-medicine applications (collaboration with Clearwater Clinical)
Real-time monitoring for vibration/acceleration, sound, air pressure, and temperature during emergency patient transport to the NICU (collaboration with the CHEO patient transport team)
Semi-supervised and species-specific prediction of microRNA from genomic sequence or transcriptomics data
Prediction of protein-protein interactions from sequence
Development of various novel assistive devices for persons who are disabled and the elderly
Identification of post-translational modifications in proteins, including methylation, sumoylation, glycosylation, and hydroxylation.