BC Platforms, RIKEN and the Finnish Institute for Health and Welfare (THL) collaborate on artificial intelligence approach to identify people most at risk of from COVID-19

RIKEN’s precision model algorithms designed to predict COVID-19 outcomes using THL’s patient data and BCP’s BC|INSIGHT

RIKEN is Japan’s largest comprehensive research institution encompassing a network of world class research centres and institutes across Japan

BC Platforms, a global leader in clinical and genomic data management, analytics and access, announces that it has joined forces with Japan’s RIKEN research institution and the Finnish Institute for Health and Welfare (‘THL’), in an international research effort to support development of a precision prediction model to identify those most at risk from COVID-19. RIKEN and THL play significant roles in Japan and Finland, respectively, in the fight against COVID-19, improving society’s resilience to the coronavirus pandemic.

RIKEN has developed precise disease stratification algorithms using information geometry and artificial intelligence which is applicable to determine which people are at risk of mortality and aggravation from COVID-19. THL leads a nationwide research project, COVIDprog, which collects samples and data from patients with positive COVID test results to support the identification of individual characteristics, including underlying genetic causes, which affect outcomes. BC-Platforms is providing its BC|INSIGHT platform to enable the curation, integration and analysis of THL’s clinical data of 300 to 1,000 Finnish COVIDprog research subjects using RIKEN’s algorithms. This is to identify a ‘COVID-19 prediction procedure’ that can estimate symptoms and outcomes of COVID-19 based on health records of infected patients obtained before SARS-CoV-2 infection.

Dr Kazuhiro Sakurada, Group Director at RIKEN, said, “SARS-CoV-2 infection exhibits inter-individual clinical variability, ranging from asymptomatic virus carriers to individuals with severe disease that rapidly progress to respiratory failure. There are known and unknown factors that influence the symptoms and outcomes of COVID-19. In addition, heterogeneity in susceptibility to infection and severity of symptoms may be present among vaccinated individuals. At RIKEN we are developing a personalized phenotypic prediction model which is applicable to identify those at high risk, potentially saving life by providing care to those groups.”

Professor Markus Perola, leader of the COVIDprog research at THL, said, “We are collaborating with RIKEN and BC Platforms to research a ‘COVID-19 prediction procedure’, making use of the rich patient data in our COVIDprog, coupled with RIKEN’s AI modelling skills and BC Platforms’ secure data management capabilities. Our hope is, together, we will gain insights into exactly who is at risk from COVID-19, and why, at a molecular/genetic level, so society can better allocate resources to those who most need them to survive the pandemic.”

Tero Silvola, CEO of BC Platforms, said, “Systematic data sharing in a safe and secure manner is crucial in our fight against COVID-19. We are delighted to collaborate with RIKEN and THL on their novel data-led approach to protecting people in the pandemic. We will use our BC|INSIGHT platform as a data-backbone, to provide the Japanese and Finnish researchers with the high-quality data they need to determine predictive procedures to support shielding and treatment of people during this, and future, pandemics. We are looking forward to working with RIKEN to make this AI prediction platform available to other healthcare systems in order to help them to fight the COVID-19 pandemic.”

Susceptibility and resistance to SARS-CoV-2 depends on multiple factors, including the central role played by innate and acquired immunity, which are in turn influenced by previous vaccinations, such as BCG, and infections from other corona viruses. The gut microbiota also helps shape the immune system, which can be damaged by inappropriate use of antibiotics. Why some people rapidly progress to respiratory failure in COVID-19 is poorly understood, with multiple associations observed between outcomes and hypertension, diabetes, obesity and cardiovascular disease. Several genes have been identified from research on other coronaviruses, including ACE2 and TMPRSS2, which are directly involved in viral infection. This international collaboration aims to unpick these phenotypic and genetic associations to provide a better understanding of specifically why and how COVID-19 leads to mortality in some patients.