91亚色 U led study found clear vaccine-initiated immune response biomarkers between HIV positive and HIV negative groups, but outliers underscore varied, intricate nature of the immune system

How people with compromised immune systems respond to vaccines is an important area of immunological research. A led by 91亚色 found that not only could machine-learning models accurately pinpoint differences in healthy controls and those living with HIV, but also found outliers in both groups that provide fascinating glimpses into the complex nature of the immune system and what personalized medicine could look like in the future, accounting for variables such as age, comorbidities and genetics.

鈥淭his study constitutes an important step forward in the potential for personal vaccination intervention strategies,鈥� says lead author Chapin Korosec, who worked on this paper as a postdoctoral fellow at 91亚色 under the supervision of Faculty of Science Professor , whose research focuses on infectious disease modelling. 鈥淏y learning the structure of immune variability at scale, we move toward a data-driven foundation for personalized vaccination and therapeutic design.鈥�

Korosec, now an adjunct professor with the University of Guelph, used a dataset of people with and without HIV who had received up to five doses of COVID-19 vaccine over the course of 100 weeks. All the individuals living with HIV were from the Greater Toronto Area whose illness was being controlled with antiretroviral therapy. The researchers used a type of machine-learning method called random forest to analyze 64 immune biomarkers elicited through a response to the COVID-19 vaccine, and then created a group of 鈥榲irtual patients鈥� to further model immune responses.

鈥淲hile we were working with a rich dataset well suited for statistical testing, longitudinal mathematical models still face identifiability limits when the data cannot uniquely resolve immune dynamics. We therefore turned to machine learning to identify the core differences between groups, and then leveraged that learned structure to generate virtual patients that capture how immune patterns differ between groups.鈥�

They were able to show that saliva-based antibodies, particularly a type of antibody in the saliva called IgA, coupled with white blood cells, which have long been known to be associated with HIV status, create the signature difference between the two groups. Korosec says this is significant because there is a lot of research showing altered mucosal immunity for those living with HIV and how it is influenced in the short and long term.

Heffernan notes that they identified subgroups within the HIV positive group, which highlights the importance of personalized vaccination strategies and the challenges of modelling immune responses due to individual variability.

鈥淭he immune response is very, very complicated.鈥� explains Heffernan. 鈥淪ometimes something can act as an inhibitor of an arm of the immune response, but in other times it might be an activator. There is also a lot of individual variability among people with similar immune system status. Using machine learning, mechanistic modelling, and 鈥榲irtual patients鈥� we can try to uncover important differences in the subgroups and between individuals 鈥� even of immune system components that are not measured in the data. Kind of like trying to find the needle in a haystack, but with a clearer path to finding it.鈥�

The HIV positive group, despite having the benefits of antiretroviral therapy, had clear differences in their vaccine-elicited responses compared to the control group and the machine-learning model was able to classify those differences with nearly 100 per cent accuracy, but there were two individuals who they could not differentiate from the control group.

鈥淣o matter how we shuffled the data or which biomarkers we used, the machine-learning algorithm could not distinguish a small subset of HIV-positive individuals from those who were HIV-negative,鈥� says Korosec. 鈥淚n those individuals, the vaccine-induced immune responses were indistinguishable from the HIV-negative group. That suggests that, at least in terms of vaccination response, their immune function was effectively restored.鈥�

Conversely, there was one individual in the healthy control group whose markers looked indistinguishable from someone living with HIV, which may suggest underlying immune issues that may not yet have been clinically identified.

Supported by the National Research Council of Canada (NRC)-Fields Mathematical Sciences Collaboration Centre, the National Sciences and Engineering and Research Council of Canada and Artificial Intelligence for Public Health (AI4PH), the study was published today as a pre-print in the Journal Patterns and will appear in print as the cover article on March 13. Korosec worked with collaborators, including Heffernan,  Senior Research Officer Mohammad Sajjad Ghaemi from the NRC Digital Technologies Research Centre, Associate Professor Jessica Conway from Pennsylvania State University and researchers from the University of Toronto and St. Michael鈥檚 Hospital.

鈥淭his study moves us closer to understanding immune diversity in people living with HIV; how their responses compare to age-matched controls, how well antibodies are maintained over time, and why some individuals show strikingly different patterns,鈥� says Korosec.

-30-

91亚色 is a modern, multi-campus, urban university located in Toronto, Ontario. Backed by a diverse group of students, faculty, staff, alumni and partners, we bring a uniquely global perspective to help solve societal challenges, drive positive change, and prepare our students for meaningful life and career paths. 91亚色's Glendon Campus is home to Southern Ontario's Centre of Excellence for French Language and Bilingual Postsecondary Education. 91亚色鈥檚 campus in Costa Rica offers students exceptional transnational learning opportunities and innovative programs. Together, we can make things right for our communities, our planet, and our future.

Media Contact: Emina Gamulin, 91亚色 Media Relations, 437-217-6362, egamulin@yorku.ca

The post Machine-learning immune-system analysis study may hold clues to personalized medicine appeared first on News@91亚色.

TORONTO, June 4, 2025 鈥� If data used to train artificial intelligence models for medical applications, such as hospitals across the Greater Toronto Area, differs from the real-world data, it could lead to patient harm. A new study out today from 91亚色 found proactive, continual and transfer learning strategies for AI models to be key in mitigating data shifts and subsequent harms.

To determine the effect of data shifts, the team built and evaluated an early warning system to predict the risk of in-hospital patient mortality and enhance the triaging of patients at seven large hospitals in the Greater Toronto Area.

The study used GEMINI, Canada鈥檚 largest hospital data sharing network, to assess the impact of data shifts and biases on clinical diagnoses, demographics, sex, age, hospital type, where patients were transferred from, such as an acute care institution or nursing home, and time of admittance. It included 143,049 patient encounters, such as lab results, transfusions, imaging reports and administrative features.

鈥淎s the use of AI in hospitals increases to predict anything from mortality and length of stay to sepsis and the occurrence of disease diagnoses, there is a greater need to ensure they work as predicted and don鈥檛 cause harm,鈥� says senior author 91亚色 Assistant Professor of 91亚色鈥檚 School of Health Policy and Management, Faculty of Health, a member of Connected Minds and a faculty affiliate at the聽Vector Institute.

鈥淏uilding reliable and robust machine learning models, however, has proven difficult as data changes over time creating system unreliability.鈥�

The data to train clinical AI models for hospitals and other health-care settings need to accurately reflect the variability of patients, diseases and medical practices, she adds. Without that, the model could develop irrelevant or harmful predictions, and even inaccurate diagnoses. Differences in patient subpopulations, staffing, resources, as well as unforeseen changes to policy or behaviour, differing health-care practices between hospitals or an unexpected pandemic, can also cause these potential data shifts.

鈥淲e found significant shifts in data between model training and real-life applications, including changes in demographics, hospital types, admission sources, and critical laboratory assays,鈥� says first author Vallijah Subasri, AI scientist at University Health Network. 鈥淲e also found harmful data shifts when models trained on community hospital patient visits were transferred to academic hospitals, but not the reverse.鈥�

To mitigate these potentially harmful data shifts, the researchers used a transfer learning strategies, which allowed the model to store knowledge gained from learning one domain and apply it to a different but related domain and continual learning strategies where the AI model is updated using a continual stream of data in a sequential manner in response to drift-triggered alarms.

Although machine learning models usually remain locked once approved for use, the researchers found models specific to hospital type which leverage transfer learning, performed better than models that use all available hospitals.

Using drift-triggered continual learning helped prevent harmful data shifts due to the COVID-19 pandemic and improved model performance over time.

Depending on the data it was trained on, the AI model could also have a propensity for certain biases leading to unfair or discriminatory outcomes for some patient groups. 

鈥淲e demonstrate how to detect these data shifts, assess whether they negatively impact AI model performance, and propose strategies to mitigate their effects. We show there is a practical pathway from promise to practice, bridging the gap between the potential of AI in health and the realities of deploying and sustaining it in real-world clinical environments,鈥� says Dolatabadi.

The study is a crucial step towards the deployment of clinical AI models as it provides strategies and workflows to ensure the safety and efficacy of these models in real-world settings.

鈥淭hese findings indicate that a proactive, label-agnostic monitoring pipeline incorporating transfer and continual learning can detect and mitigate harmful data shifts in Toronto鈥檚 general internal medicine population, ensuring robust and equitable clinical AI deployment,鈥� says Subasri.

The paper, , was published today in the journal JAMA Network Open.

The post New research finds specific learning strategies can enhance AI model effectiveness in hospitals appeared first on News@91亚色.

TORONTO, May 5, 2025 鈥� Getting a timely diagnosis of autism spectrum disorder is a major challenge, but new research out of 91亚色 shows that how young adults, and potentially children, grasp objects could offer a simpler way to diagnose someone on the autism spectrum.

The team, part of an international collaboration, used machine learning to analyze naturalistic hand movements 鈥� specifically, finger motions during grasping 鈥� in autistic and non-autistic individuals.

鈥淥ur models were able to classify autism with approximately 85 per cent accuracy, suggesting this approach could potentially offer simpler, scalable tools for diagnosis,鈥� says lead author, Associate Professor of 91亚色鈥檚 Department of Psychology and the Centre for Vision Research.

鈥淎utism currently affects about one in 50 Canadian children, and timely, accessible diagnosis remains a major challenge. Our findings add to the growing body of research suggesting that subtle motor patterns may provide valuable diagnostic signals 鈥� something not yet widely leveraged in clinical practice.鈥�

In addition to social and communication challenges, autism, a neurodevelopmental disorder, can include motor abnormalities which often show up in early childhood. The researchers say testing for these motor movements early could lead to faster diagnoses and intervention.

鈥淭he main behaviours markers for diagnosis are focused on those with relatively late onset and the motor markers that can be captured very early in childhood may thus lower age of diagnosis,鈥� says Professor of the University of Haifa, an expert in autism research and a key collaborator in this study.

Autistic and non-autistic young adult participants were asked to use their thumbs and index fingers, which had tracking markers attached, to grasp different blocks of varying size, lift each one and replace it in the same spot, and put their hand back in the starting position. The researchers used machine learning to analyze the participants鈥� finger movements as they made grasping motions.

Both groups of participants had normal IQ and were matched on age and intelligence. Young adults were used instead of children to rule out any differences in the findings due to delayed development.

The research found that subtle motor control differences can be captured effectively with more than 84 per cent accuracy. The study also showed there were distinct kinematic properties in the grasping movements between autistic and non-autistic participants.

Analysis of naturalistic precision grasping tasks has not typically been used in previous studies, says Freud. Machine learning, however, provides researchers with a powerful new tool to analyze motor patterns, opening new ways to use movement data in the assessment of autism spectrum disorder.

The findings, says Freud, could lead to the development of more accessible and reliable diagnostic tools as well as timely intervention and support that could improve outcomes for autistic individuals in the future.

The paper, , was published today in the journal Autism Research.

The post How individuals grasp an object may offer simpler diagnosis for autism spectrum disorder appeared first on News@91亚色.

TORONTO, May 26, 2021 鈥� What level of immunity against COVID-19 do you have after being vaccinated or contracting the virus? 91亚色 Professor is receiving a $200,000, one-year grant from the National Research Council of Canada (NRC) to understand the rate of immunity in both of these scenarios.

The is part of the NRC鈥檚 designed to bring the best Canadian and international researchers together to fast track research and development aimed at specific COVID-19 gaps and challenges as identified by Canada's health experts.

Heffernan, Inaugural 91亚色 Research Chair (Tier II), Multi-Scale Methods for Evidence-based Health Policy in the Faculty of Science, is leading the study with colleagues James Ooi, the NRC鈥檚 Pandemic Response Challenge program project lead, and M. Sajjad Ghaemi, NRC research officer, both from the NRC-Fields Collaboration Centre.

鈥淒ifferent vaccines elicit an immune response using different pathways, which affects the level and type of immunity you build,鈥� says Heffernan of the Canadian Centre for Disease Modelling. 鈥淲ith this research, we鈥檙e tracking the activation of the immune response that鈥檚 been excited by vaccines, looking at the generation of antibodies, as well as memory B cells and T cells. Clinical trials can measure the number of antibodies, but they don鈥檛 measure B cells and T cells.鈥�

To do this, the researchers will combine mathematical models of immunity development (mechanistic models) with machine learning algorithms to better understand the outcomes of immunity to the SARS-CoV-2 virus after one- and two-dose regimes of adenovirus (AstraZeneca, Johnson & Johnson), mRNA (Pfizer and Moderna) and protein subunit (Novavax) vaccines. They will model the effectiveness and immunity response to the virus, including pathogen mutations and variants, when vaccines doses are given days or weeks apart or as is the case in Canada currently, four months apart.

The researchers hope the mechanistic models will enrich the dataset upon which the machine learning framework is trained. By combining new datasets that are being released publicly, this approach can potentially advance the accuracy of the machine learning framework. This will allow the researchers to classify outcomes of vaccinations as emerging evidence becomes available.

The idea is to uncover the complex interactions between interferon signalling pathways and the adaptive immune response to SARS-CoV-2 infection and vaccination.

鈥淲hen you model outcomes in antibodies, it鈥檚 important to try to model the development of these memory cells in the background. Antibodies protect you from being infected and if they fail, it鈥檚 the memory cells that give you that activate factor that allows you to have a milder infection,鈥� says Heffernan.

One of the goals of this research is to tailor vaccines to people鈥檚 body chemistry. 鈥淭his is well into the future, but the goal eventually is to develop inhouse models for mRNA, adenovirus and protein subunit vaccines that can be used to inform what type of vaccine a person should get depending on the characteristics of their immune system,鈥� says Heffernan.

In the short-term, the researchers hope to predict the outcomes in children of various vaccines even without the results of a clinical trial. Based on the differences in immune response of children versus adults, the idea is to change the machine learning and mechanistic models calibrated for adults so that they fit the characteristics of children.

The modelling can also be expanded in the future to test other types of vaccines for COVID-19, in addition to vaccines for other viruses.

The data will be provided to public health agencies, such as the Public Health Agency of Canada, the National Advisory Committee on Immunization, the Canadian Immunization Research Network, and academic researchers to inform vaccine design and policy, and predict safety and efficacy of different vaccine types.

-30-

91亚色聽is a modern, multi-campus, urban university located in Toronto, Ontario. Backed by a diverse group of students, faculty, staff, alumni and partners, we bring a uniquely global perspective to help solve societal challenges, drive positive change and prepare our students for success. 91亚色's fully bilingual Glendon Campus is home to Southern Ontario's Centre of Excellence for French Language and Bilingual Postsecondary Education. 91亚色鈥檚 campuses in Costa Rica and India offer students exceptional transnational learning opportunities and innovative programs. Together, we can make things right for our communities, our planet, and our future.聽

Media Contact:

Sandra McLean, 91亚色 Media Relations, 416-272-6317,聽sandramc@yorku.ca

The post How immune are you after one or two doses of a COVID-19 vaccine? appeared first on News@91亚色.

91亚色 expert predicts how artificial intelligence (AI) will improve your life

TORONTO, December 16, 2019 鈥� The future of AI could mean going to your local caf茅 only to be greeted by a robot or getting a facial recognition scan at the airport instead of showing your passport, according to Karthik Kuber, an artificial intelligence expert in 91亚色鈥檚 School of Continuing Studies.

An instructor in 91亚色鈥檚 certificate program, Kuber predicts AI will transform the way we work, shop, travel and play in 2020 and beyond. That鈥檚 because we鈥檒l rely on technology more to solve human problems and make life easier.

Some GTA residents, for example, will use AI-driven apps on their phones to fight parking tickets, he said. Others will use AI-equipped smart shopping carts and apps to scan their grocery items, track their bill and pay it. Kuber also pointed to Beijing鈥檚 new international airport where passengers use facial recognition technology to check in and pass security, without the need for identification or passports. The future will also include more robots performing simple tasks, like greeting customers, and advanced AI software that will defeat the world鈥檚 best gamers including chess players.

But does that mean machines will replace human workers? No, says Kuber, who cites the example of ATM machines first being introduced to the marketplace, which resulted in more bank branches, more human tellers and reduced branch operating costs.

鈥淢ore jobs will be created by automation than lost,鈥� explains Kuber. 鈥淭here will always be a need for traffic cops pulling over speeders, for example, but a future with more traffic cameras everywhere means more data to analyze and more jobs in computer science and data analysis.鈥�

Kuber has extensive AI and machine learning experience working in the technology and banking sectors, and a PhD in computer science, focused on machine learning and evolutionary computation. He is a co-organizer of the and is an active volunteer with , which helps organizations use data science for social causes.

He can comment on:

• How artificial intelligence (AI) will make life easier
• Predictions for robotic process automation
• How machine learning techniques will solve real-world problems
• The evolution of computer algorithms to improve efficiencies
• How AI experts are creating future social networks and social media networks

91亚色 champions new ways of thinking that drive teaching and research excellence. Our students receive the education they need to create big ideas that make an impact on the world. Meaningful and sometimes unexpected careers result from cross-disciplinary programming, innovative course design and diverse experiential learning opportunities. 91亚色 students and graduates push limits, achieve goals and find solutions to the world鈥檚 most pressing social challenges, empowered by a strong community that opens minds. 91亚色 U is an internationally recognized research university 鈥� our 11 faculties and 25 research centres have partnerships with 200+ leading universities worldwide. Located in Toronto, 91亚色 is the third largest university in Canada, with a strong community of 53,000 students, 7,000 faculty and administrative staff, and more than 300,000 alumni.

91亚色 U's fully bilingual Glendon Campus is home to Southern Ontario's Centre of Excellence for French Language and Bilingual Postsecondary Education.

Media Contact: Vanessa Thompson, 91亚色 Media Relations, 647-654-9452,聽vthomps@yorku.ca

The post Will robots take over the world in 2020? appeared first on News@91亚色.