Right now, we hear a lot about Delta and Omicron, the two most circulating variants in France, but little do we know that more than 70,000 variants were discovered between early October 2021 and late November 2021. BioNTech and InstaDeep announce that they have developed and successfully tested an evolutionary alert system to detect high-risk variants of SARS-CoV-2 (Covid-19). This combines publicly available SARS-CoV-2 sequence information with predictive AI analysis.

Since the emergence of Covid-19, research centers have been mobilizing to find vaccines or drugs that could stop the disease. The exponential number of variants makes their task difficult, as does classifying them according to their dangerousness. BioNTech and InstaDeep announced on January 14 the development of a new method combining structural modeling of the Spike protein with artificial intelligence to identify and monitor in real time the appearance and evolution of high-risk variants of SARS-CoV-2, called the Early Warning System (EWS). This warning system predicts immune evasion and infectious potential of variants using artificial intelligence.

BioNTech

Founded in 2008 in Mainz, Germany, BioNtech (Biopharmaceutical New Technologies) is a next generation immunotherapy company. A pioneer in new therapies for cancer (including mRNA-based therapies) and other serious diseases, BioNtech’s ambition is to become the world’s leading biotech company for individualized cancer medicine.
With expertise in mRNA vaccine development and in-house manufacturing capabilities, BioNTech and its collaborators are developing several mRNA vaccine candidates for a range of infectious diseases (e.g. BioNTech-Pfizer Covid-19).

InstaDeep

Founded in 2014 by Karim Beguir and Zohra Slim, InstaDeep is a leader in decision AI systems, having been named two years in a row to the CB Insights AI 100 ranking of the 100 most
companies in the world.
The company develops patented AI products such as its DeepChainTM protein design platform that was used in the design of EWS and collaborates with leading companies such as Google DeepMind, Nvidia and Intel. To contribute to the development of AI, it develops solutions for everyone, published in open source, and organizes events and training.

Covid-19 mutations

Identifying high-risk variants requires considerable laboratory testing time. EWS enables early detection of these variants and reduces the time needed by health authorities to assess their impact and react in a timely manner.

Like any virus, Covid-19 experiences new mutations in its genome, which can change the characteristics of the original strain identified in early 2020. More than 13,400 non-synonymous mutations (which alter the protein sequence) have been observed in the Spike protein alone. Available data show that thousands of new variants are appearing weekly at an increasing rate, with a weekly average of variants recorded of approximately 300 in September 2020, 7,000 in August 2021, and 12,000 in December 2021.
While most mutations are unfavorable to the virus or have no impact on its characteristics, some individual mutations or combinations of mutations lead to high-risk variants (HRVs) with altered immune evasion capabilities and/or enhanced transmissibility. A variant capable of bypassing antibody neutralization is of particular importance and poses a risk to previously infected individuals with COVID-19 and those who are fully vaccinated.
BioNTech and InstaDeep have been working since late 2020 to design an early warning system that can identify potentially high-risk Covid-19 variants as early as possible.

Early Warning System (EWS), the alert system developed by InstaDeep and BioNTech

This new computational method combines structural modeling of the Spike viral protein with artificial intelligence algorithms to flag potential high-risk variants within 24 hours, based on parameters assessing their infectious potential (e.g., the interaction between the ACE2 protein and the Spike protein variant) and their risk of immune evasion. The companies validated these predictions using internally generated experimental data and publicly available data.

Identification of more than 90% of variants

In the various tests performed, EWS demonstrated its ability to identify more than 90% of the variants designated by the World Health Organization (Variants of Concern, VOC; Variants of Interest, VOI; and Variants Under Surveillance, VUM) with an average of two months’ advance. The WHO-designated Alpha, Beta, Gamma, Theta, Eta, and Omicron variants were detected by EWS the same week their genetic sequence was shared on international databases. The Omicron variant was classified as a high-risk variant by EWS on the same day that its genetic sequence was published. EWS also assessed the IHU variant as being of less concern based on current data.

The results of the study demonstrate that EWS is capable of evaluating new variants within minutes and monitoring mutant lines in near real time.
The system automatically evolves in real time as new variant data becomes publicly available.

Ugur Sahin, M.D., CEO and co-founder of BioNTech states:

“Using the advanced computational methods we have developed over the past few months, we can analyze the various Spike protein sequences and classify novel variants based on predictions of their immune evasion score and affinity for the ACE2 protein. Early identification of potentially high-risk variants could be an effective tool to alert researchers, vaccine developers, health authorities and policy makers, and give them more time to formulate a response to new variants of concern.”

Karim Beguir, co-founder and CEO of InstaDeep on his part:

“There are more than 10,000 new variant sequences being discovered every week and the scientific community simply cannot cope with so much complex data to process anymore. InstaDeep has met this challenge by deploying the powerful computational capabilities of its DeepChain artificial intelligence platform, combined with BioNTech’s SARSCoV-2 expertise and technology. For the first time, high-risk variants could be detected in a very short period of time, saving valuable months of research for the entire scientific community. We are pleased to make our work available to the scientific community and, more importantly, look forward to its tangible impact around the world.”

Translated from Covid-19 : Focus sur Early Warning System (EWS) développé par BioNTech et InstaDeep pour détecter les variants à haut risque du SARS-CoV-2