��ݮ��Ƶ

A trio of recent ��ݮ��Ƶ graduates has tackled a long-standing problem in clinical research with an automated solution that could help scientists analyze blood stem cells faster and more accurately.

“During my co-op placement at Princess Margaret Cancer Centre I was working on a project that used immunofluorescence image analysis to answer a specific scientific question about hematopoietic stem cells,” said Isabella Di Biasio, a recent graduate of ��ݮ��Ƶ’s biochemistry program. “I was having difficulties related to the ImageJ software researchers use. In talking to my supervisor, , I learned that many people at the cancer centre shared similar frustrations.”

L to R: Veronika Sustrova, Katarina Makivic, Isabella Di Biasio

Veronika Sustrova graduated with a BCS in spring 2025. Through previous internships, projects and hackathons, she has gained experience in UI development and software engineering. She is fascinated by the creative possibilities tech offers, particularly how tech can be used for social good. Veronika will begin her full-time role as a software developer at AWS DynamoDB in fall 2025.

Katarina Makivic graduated with a BCS in spring 2025. She is a passionate software developer with experience in backend development and proficiency in Python, Java, C++, Go and C. With strong problem-solving skills and industry experience in designing and implementing scalable applications, she tackles complex challenges and fosters innovation. Katarina will begin her full-time role as a software developer at AWS DynamoDB in fall 2025.

Isabella Di Biasio graduated with a BSc in biochemistry in spring 2025. Her work experiences include being a research student at the Hospital for Sick Children. More recently, she worked with Dr. Stephanie Xie, with whom she will continue her studies this fall as a doctoral student in the Department of Medical Biophysics at the University of Toronto. Isabella’s HSC research was done in collaboration with ’s lab at Princess Margaret Cancer Centre.

Hematopoietic stem cells, or HSCs, are found in bone marrow and are responsible for generating all blood cells throughout a person’s life. They give rise to all three types of blood cells — red blood cells, which transport oxygen to tissues and remove carbon dioxide; white blood cells, which are essential for immune defence; and platelets, which help blood clot and prevent excessive bleeding.

“One of the remarkable properties of hematopoietic stem cells is their ability to self-renew, meaning they can replicate themselves to maintain the pool of stem cells,” Isabella explained. “They can also remain in a quiescent state for extended periods. These cells are central to maintaining blood and immune system health. But if they don’t function as they should, they can contribute to immunodeficiency diseases and cancers of the blood such as acute myeloid leukemia.”

To study HSCs, researchers typically use bioimage analysis, a method that involves analyzing microscope images of cells to extract quantitative information. Researchers measure fluorescence intensity density, a metric that captures the brightness and concentration of fluorescent markers bound to specific proteins in the cells. By examining the fluorescence intensity of these biomarkers, scientists can assess a blood cell’s “stemness” — its ability to remain a stem cell versus one that has begun to differentiate into a type of blood cell.

“The process is incredibly manual,” Isabella explains. “Scientists use a software tool called ImageJ to classify cells, but they have to write their own scripts and analyze hundreds, even thousands of images by hand. It’s time consuming and prone to error.”

Recognizing the need for a solution, Isabella began to think how automation and machine learning might help, but as a biochemistry student she didn’t have the specific background in computer science.

That’s when she teamed up with Katarina Makivic and Veronika Sustrova, two recent graduates from ��ݮ��Ƶ’s Computer Science program. The three students formed a team through the Interdisciplinary Capstone Design Course, an initiative that brings together students from across the university’s six faculties to tackle real-world challenges.

“Katarina and I knew each other from a previous co-op placement and she told me about Isabella’s work,” Veronika recalled. “I had taken a few biology electives that same term, so this seemed like a great opportunity to apply my computer science knowledge to a practical problem involving human cells. I was also excited by the potential of creating a project that could greatly improve scientists’ scripting experience and quality of life at work.”

“Once we were introduced to each other and to the project we began to have meetings, usually weekly,” Katarina added. “We’d hash out the biology and computer science terminology and figure out what was relevant to the project. We were all interested in participating in this intersection of skills and knowledge. That’s what drew us in — the possibilities that exist if we combine our fields.”

Bringing together their expertise in biology and computer science, the trio developed a two-part solution. The first component they created was a domain-specific language designed for HSC analysis to allow researchers to write and customize image-processing scripts more easily. The second was developing a machine learning–based classifier to automatically identify HSCs in microscope images by analyzing the expression of key biomarkers.

The project was developed with researchers at Princess Margaret Cancer Centre and focused on user experience and practical design. It is the first to combine both co-op and capstone components under a recent Memorandum of Understanding between ��ݮ��Ƶ and Princess Margaret Cancer Centre, a part of the University Health Network. This partnership aims to advance cancer research and address urgent health care challenges. Enabling students to explore real-world problems during co-op placements, then develop solutions through their capstone projects, sparks new innovations and enriches education.

“Our goal was to make something the scientists could use in the lab, not just a research prototype,” Isabella said.

By streamlining image analysis and automating key steps, the trio’s tool aims to accelerate scientific discovery in stem cell biology and possibly contribute to the development of new clinical treatments.

“We’re hoping our tool will free up their time and mental space so the researchers can focus on the science,” Katarina said. “And we’re hoping that as we further refine our machine learning model, that it will be able to detect patterns in the image data and make predictions — finding something that’s invisible to the human eye, perhaps leading to new insights and discoveries.”

The team presented their work at the 2025 Interdisciplinary Capstone Design Symposium, an event that brought together undergraduate students from all six faculties at ��ݮ��Ƶ. Organized in collaboration with the Future Cities Institute, the symposium featured 34 student-led projects developed in partnership with organizations. A summary of the team’s research — abstract number 25 — is presented in the symposium’s booklet.