From a near-infinite set of possibilities, we can identify the variants that present the greatest threats.
Viral outbreaks are inherently unpredictable.
Many of the world’s most threatening viruses continue to evolve faster than new vaccines can be developed.
Instead of chasing viruses as they evolve, we are working to get ahead of them. That's why we built Octavia™.
Octavia
[ ok-tey-vee-uh ] noun
Meet Octavia™
Octavia is our biology-enabled AI platform.
Octavia generates landscape maps of the existing and potential variants of a virus based on their ability to bind human cells and evade antibodies. Using these maps, we can define the combination of antibodies (antibody repertoire) that would fill gaps in our defenses with new vaccines that protect against current and potential threats.
Unlike fully computational approaches, our platform is designed from the bottom-up to seamlessly integrate experimental insights and artificial intelligence.
Octavia relies on several enabling technologies like protein display systems, deep mutational scanning, single-cell sorting, and deep sequencing.
Here's how it works:
Apriori isolates a key protein fragment from the target virus
Octavia generates millions of non-replicating, non-infectious variant fragments in vitro from anchor sequences
We relocate the fragments to a display system that lets us measure how these proteins interact with antibodies as well as human receptors for viral entry into cells
We use the resulting massive protein–protein interaction data set to train our biology-enabled AI using machine learning
Our AI deduces how other potential variants might bind to different antibodies, and models how each variant would behave in real life to identify those that pose the greatest potential threat
This allows us to reverse engineer variant-resilient vaccines that confer protection against a broad range of the most concerning potential variants
