64th ISI World Statistics Congress - Ottawa, Canada

64th ISI World Statistics Congress - Ottawa, Canada

IPS 481 - Informing response to COVID19 using mathematical and machine learning modeling and analytics

Category: IPS
Wednesday 19 July 2 p.m. - 3:40 p.m. (Canada/Eastern) (Expired) Room 201

View session detail

Context: Global efforts have gone into building accurate and predictive models of different aspects of COVID-19 from the limited but ongoing flow of data available during the pandemic, in the hope of developing better diagnostic, prognostic, treatment, and public health tools, but also to better understand EDI impacts of the pandemic.    This session will present the outcome of multiple collaborations between the Digital Technology research center at the National research council of Canada, and it’s network of academic partners across the country. Presentations will cover a whole range of successful modeling efforts, from in-host models for current vaccines, to predictive models of COVID resilience and response built from molecular data, impact of COVID on EDI indicators, and new diagnostic COVID-19 screening tools.  Proposed format: 4 speakers Proposed duration: 100 minutes. This session proposal is composed of 4 presentations, i.e. 4 papers. Presentation - 1  ------------- · Presenter: Ashkan Ebadi, National Research Council of Canada · Presentation title: COVID-19 screening tools based on medical images using few-shot learning strategies · Summary: Since the beginning of 2020, the COVID-19 pandemic has had an enormous impact on the global healthcare systems, and there has not been a region or domain that has not felt its impact in one way or another. The gold standard of COVID-19 screening is the reverse transcription-polymerase chain reaction (RT-PCR) test. With RT-PCR being laborious and time-consuming, much work has gone into exploring other possible screening tools to observe abnormalities in medical images using deep neural network architectures. But, such deep neural network-based solutions require a large amount of labelled data for training. In this talk, we will first briefly introduce the few-shot learning approach in which models are built such that they can adapt to novel tasks based on small numbers of training examples. Next, we will see its application in a real-life example where we used few-shot learning strategies to build an open-source explainable model sensitive to COVID-19 positive cases, using a very limited set of annotated data. The model can generalize from a few examples by employing a unique structure to rank similarities between inputs without necessitating extensive retraining. The developed technology is low-cost and non-invasive and can adapt to new pandemics and diseases. · Collaborators o Dr. Alexander Wong – University of Waterloo o Dr. Adrian Florea – McGill University o Dr. Sonny Kohli – McMaster University o Stéphane Tremblay and Dr. Pengcheng Xi – National Research Council of Canada o And, multinational companies such as Microsoft and Redhat.   Presentation 2 ---------------  · Presenter: James Ooi, National Research Council of Canada · Presentation title: COVID-19 in-host modelling: applying mechanistic models to understand the vaccine-induced immune response. · Summary At the onset of the COVID-19 pandemic, several vaccine candidates (adenovirus vector, mRNA and protein subunit vaccine) approvals were accelerated for emergency use. The approval was guided by limited clinical trial data performed within a short study time frame. Due to these unprecedented data challenges, there exist gaps in understanding various aspects of these vaccines, including but not limited to host immunogenicity beyond the study time frame, varying doses, age, sex and effects of adjuvants. By applying limited published clinical trial data, we developed within-host mathematical models for various vaccine types that are currently in use or in the final stages of clinical trials in Canada and analyse its associated humoral and cellular adaptive immune responses. The vaccine-induced immune response investigated includes the antibodies, T helper cells, cytokines and cytotoxic T lymphocytes. The models’ prediction allows for a better understanding of the relationship between immune cells and cytokines while parameter sensitivity analysis establishes the factors that contribute to peak immune response of different vaccine types. The long-term antibody prediction shows a discernible degradation. This finding supports the current third-dose booster guidelines. Our within-host models guide the vaccination strategy by health authorities to optimize the vaccine rollout and serve as an in-silico tool for future vaccine re-formulation.  · Collaborators: o Prof. Jane Heffernan - York University o Dr. Mohammad Sajjad Ghaemi - NRC-Fields Collaboration Centre o Prof. Mario Ostrowski - University of Toronto o Prof. James Watmough - University of New Brunswick o Prof. Morgan Craig - University of Montreal.  Presentation - 3 -------------------- · Presenter: Miroslava Čuperlović-Culf, National Research Council of Canada · Presentation title: Machine learning modeling of the effect of COVID19 on metabolic network. · Summary Models of metabo-lipidomics network obtained from high throughput omics measurements and developed using hybrid data and knowledge based approaches show alteration in the metabolic steady state in severe cases of COVID19. Sex and age effects on this change are explored using Gaussian Process Regression, a Machine learning approach as well as distance correlation, statistical method and related to possible long term effects of COVID19. These novel approaches, augmented with univariate analysis of changes in lipidome and metabolome, provide information about functional systemic changes that identify potential biomarkers, novel therapeutic targets for more effective anti-viral treatments or adjuvant targets to augment vaccination strategies. · Collaborators: o Dr. Miroslava Čuperlović-Culf - National Research Council of Canada o Dr. Irina Alecu - University of Ottawa o Anuradha Surendra - National Research Council of Canada o Thao Nguyen-Tran - University of Ottawa o Dr. Angela M. Crawley - Ottawa Hospital Research Institute o Dr. Michaeline McGuinty - Ottawa Hospital Research Institute o Dr. Steffany A.L. Bennett - University of Ottawa.  Presentation - 4 ------------------ · Presenter: Steffany A.L. Bennett, University of Ottawa · Presentation title: Impact of COVID-19 on critical indicators of academic success in context of equity, diversity and inclusion at the University of Ottawa using big data/machine learning approaches. · Summary Critical indicators of academic success in context of equity, diversity, and inclusion initiatives and as influenced by response to the COVID-19 pandemic at the University of Ottawa are explored using Gaussian Process Regression, distance correlations, and univariate analysis visualized and communicated to stakeholders using dynamic real-time data dashboards. We present real-world application of big data/machine learning approaches to contextualize challenges and opportunities influencing academic success impacted as a result of the COVID-19 pandemic. · Collaborators: o Bensun Fong - University of Ottawa o Miroslava Cuperlovic-Culf - National Research Council of Canada o Steffany A.L. Bennett - University of Ottawa

 

Global efforts have gone into building accurate and predictive models of different aspects of COVID-19 from the limited but ongoing flow of data available during the pandemic, in the hope of developing better diagnostic, prognostic, treatment, and public health tools, but also to better understand equity, diversity, inclusion impacts of the pandemic.  
 
This session will present the outcome of multiple collaborations between the Digital Technology research center at the National research council of Canada, and it’s network of academic partners across the country. Presentations will cover a wide range of successful modeling efforts, from in-host models for current vaccines that help to understand the vaccine-induced immune response, to predictive models of COVID resilience and response built from molecular data, impact of COVID on equity, diversity and inclusion indicators, and new diagnostic COVID-19 screening tools based on medical images, using few-shot learning strategies.

 

Organiser: Sevgui Erman 

Chair: Louis Borgeat  

Speaker: Ashkan Ebadi 

Speaker: Miroslava Čuperlović-Culf  

Speaker: James Ooi 

Speaker: Steffany Bennett 

Good to know

This conference is currently not open for registrations or submissions.