Nec Corp.’s bioinformatics subsidiary Nec Oncoimmunity AS said it is working with Oslo University Hospital to develop an artificial intelligence (AI) platform that will allow the team to design a T-cell diagnostic to complement the current serological tests for infectious diseases, including COVID-19.
The aim of the project is to improve the ability to identify immune responses and acquired immunity. The AI-designed T-cell diagnostic is expected to complement antibody tests and to identify individuals who are naturally immune to the virus after being infected with SARS-CoV-2 or other seasonal coronaviruses and who have acquired immunity after vaccination.
“We have obtained results by using the conventional strategies after seven months of extensive work. We will have the first results from the new AI-based analysis within weeks of testing of the AI-predicted epitopes,” Ludvig Munthe, head of research and group leader at the department of immunology at Oslo University Hospital, spoke to BioWorld about the development timeline for the project.
“Results will be used to define epitopes that can be tested in patients to validate vaccine responses after vaccination commences and to identify high risk groups that have inadequate immune responses,” he said. “Results will also feed back to the AI, machine learning platform at Nec Oncoimmunity, where we hope that the Immune Profiler that they have developed will revolutionize the way we work on this and future epidemics.”
Munthe added that the team at the Norwegian hospital has already tested T-cell responses from more than 80 individuals using conventional epitope discovery of 115 epitopes.
“We will now be able to proceed and test several hundred new epitopes to identify optimal, so-called immunodominant epitopes that are the basis of protective immunity against SARS-CoV-2 (the COVID-19 virus). This will allow a side-by side comparison of the old and new strategies, and also allow optimized machine learning at Nec Oncoimmunity,” he added.
Oslo, Norway-based Nec Oncoimmunity said T-cell diagnostics “represent a ‘blind spot for the monitoring of immunity to COVID-19 in the world’s population,” and the diagnostic to be developed in this project will monitor the underlying T-cell response to the infection.
Under this partnership, Nec Oncoimmunity will contribute its machine learning-based software called Nec Immune Profiler. It integrates modules of human leukocyte antigen binding, processing, and antigen presentation in a system to accurately predict antigens that are naturally processed and presented to the tumor cell surface.
Nec Immune Profiler was developed to predict neoantigens from next generation sequencing data for personalized cancer immunotherapy and cancer immunotherapy biomarkers, as Nec Oncoimmunity is focused on cancer immunotherapies.
Nec Oncoimmunity’s chief scientific officer Trevor Clancy said the team will adapt and leverage the Nec Immune Profiler and other AI technologies at Nec Corp. to develop a COVID-19 T-cell diagnostic for the diverse genetic makeup in the global population.
The AI platform is expected to help contact-trace and control transmission in the fight against COVID-19. It may also be used in future emergency settings to rapidly develop diagnostics against new pandemics caused by novel infectious agents.
Nec Oncoimmunity’s CEO Richard Stratford said the AI platform will be applied first to the current COVID-19 pandemic, but stressed the use is not limited.
“We will design this platform to be future-proof and make it applicable to any future emerging infectious agent that could threaten the global population,” he said, adding that the approach will open new opportunities in the growing infectious disease diagnostics market for the company.
The diagnostic to be developed in this project will also address issues from the current technologies, which can result in extensive trial and error to define exactly which parts of the pathogen induces robust immunity.
“The main challenge today is epitope discovery, that means to identify the parts of a virus or organism that can be targeted by the immune system. Today this process is arduous and error prone. At our research center, we have spent seven months using conventional means to identify such epitopes. This is inefficient and based on extensive and expensive trial and error,” Munthe explained to BioWorld.
“Together with Nec Oncoimmunity, we will now take the important steps to change the way we work and think in such a discovery process. We have high hopes that machine learning and AI identified epitopes will revolutionize this process for this and new emerging diseases,” he added.