Unlocking the hidden mysteries within health data relies on skillfully analyzing and applying that data in innovative ways. With a myriad of data sources at researchers’ fingertips these days, the potential for advancements in health research are limitless, from early allergy detection tools to new diagnostic practices for cancer.  

But how can we unleash the power of this data? The Centre for Computational Medicine (CCM) at The Hospital for Sick Children (SickKids) is leveraging its expertise to inform health research across the country and bridge the gap between the research community and its data.

Empowering precision-based medicine through cutting-edge data science and computational power

Two teams help this facility run smoothly: CCM Bioinformatics and Research IT High Performance Computing. The High Performance Computing Facility, located in the SickKids Research Institute’s Data Centre, is home to over 30,000 CPU cores, 27 Petabytes of storage and a team dedicated to maintaining state-of-the-art hardware and equipment. Offering a range of research data analysis and storage capabilities, High-Performance Computing is the backbone of services provided by the facility and helps ensure the safety and security of confidential data from research projects across Ontario, including SickKids research.

“We are one of the few facilities in Ontario that has the expertise and infrastructure to maintain this equipment in a secure way,” explains Fernando Silva, Manager of High Performance Computing. “As more scientists explore how innovation in areas like artificial intelligence and machine learning can inform their research, there has been a massive increase in demand for the kind of high processing strength and data storage we offer.”

This is particularly true for Precision Child Health, a movement at SickKids to reimagine how we can deliver precision-based medicine using integrated data about each patient and their family. From genetic code to postal code information, the vision of Precision Child Health relies on accurate and robust data to perform meaningful analyses that are tailored to each person, and that requires both computational power and data science expertise.

“Computational power will underpin all advancements in Precision Child Health,” says Andrei Turinsky, Manager of CCM Bioinformatics. “You cannot do precision medicine without understanding how to use data science methods like machine learning for classification, diagnosis and analysis on the individual patient level.”

Providing expertise and tools to support research teams

The CCM Bioinformatics team plays a vital role as consultants for scientists by joining other labs as specialists for longer-term research projects. Whether it's hardware, software or analytics, CCM team members step in to provide valuable support, ensuring that the labs have the essential tools and capabilities to effectively carry out their research.

“Different teams and labs may have varying levels of technological expertise, and that’s okay — we are your data shop within the SickKids Research Institute,” says Turinsky. “From deciding what algorithm to use to determining what to do with your data, we can provide knowledge, expertise and tools to scientists looking for the most effective and efficient ways to advance their research.”

Their support is not limited to the SickKids community and includes external researchers and consortiums such as the pan-Canadian Care4Rare consortium, which uses genetic expertise and pipelines developed at SickKids to diagnose children with rare disease across the country. With research exploring the utility and efficacy of clinical RNA sequencing from Drs. Gregory Costain and Jim Dowling, and new data-driven methods from CCM, Care4Rare is continuing to grow and support the more than 900,000 children with rare diseases in Canada.

In addition to national and collaborative projects, CCM currently supports over 30 research projects at SickKids, including statistical methods to improve allergy diagnosis, deep-learning approaches to analyze medical imaging, and software that lets users apply machine-learning models to help identify neurodevelopmental conditions.