ADIOS: Antibody Development via Opponent Shaping

Sebastian Rene Towers, Aleksandra Kalisz, Philippe A. Robert, Alicia Higueruelo, Francesca Vianello, Chloe Ming-Han Tsai, Harrison Steel, Jakob Nicolaus Foerster

TL;DR: We present ADIOS: a method for designing antibody "shapers" which both act as effective antiviral therapies and influence the viral evolution towards less harmful variants.

Abstract:

Anti-viral therapies are typically designed to target only the current strains of a virus, a *myopic* response. However, therapy-induced selective pressures drive the emergence of new viral strains, against which the original myopic therapies are no longer effective. This evolutionary response presents an opportunity: our therapies could both *defend against and actively influence viral evolution*. This motivates our method ADIOS: Antibody Development vIa Opponent Shaping. ADIOS is a meta-learning framework where the process of antibody therapy design, the *outer loop*, accounts for the virus's adaptive response, the *inner loop*. With ADIOS, antibodies are not only robust against potential future variants, they also influence, i.e. *shape*, which future variants emerge. In line with the opponent shaping literature, we refer to our optimised antibodies as *shapers*. To demonstrate the value of ADIOS, we build a viral evolution simulator using the Absolut! framework, in which shapers successfully target both current and future viral variants, outperforming myopic antibodies. Furthermore, we show that shapers modify the distribution over viral evolutionary trajectories to result in weaker variants.
We believe that our ADIOS paradigm will facilitate the discovery of long-lived vaccines and antibody therapies while also generalising to other domains. Specifically, domains such as antimicrobial resistance, cancer treatment, and others with evolutionarily adaptive opponents. Our code is available at https://github.com/olakalisz/adios.

Lay Summary:

When scientists develop vaccines or antibody treatments, they typically design them to fight viruses as they exist today. However, viruses constantly evolve to escape our defences, making treatments less effective over time, as we saw with COVID-19 variants that reduced vaccine effectiveness.

We created ADIOS, a computer system that designs smarter antibodies called "shapers." Instead of just defending against current viruses like traditional approaches, shapers actively influence how viruses evolve, like a chess master who forces their opponent into weak positions several moves ahead. Our system simulates hundreds of generations of viral evolution to find antibodies that remain effective and actually steer viruses toward weaker forms that are easier to target.

In our computer simulations across multiple viruses, including dengue, West Nile, and influenza, shapers significantly outperformed traditional antibodies and successfully guided viral evolution toward more vulnerable variants. By using more powerful simulations, this approach could lead to longer-lasting vaccines and treatments that stay ahead of viral evolution. The same strategy could also be applied beyond infectious diseases to cancer treatment, where tumours similarly evolve to escape therapies, potentially transforming how we develop treatments for any rapidly evolving disease.

Application-Driven Machine Learning: This submission is on Application-Driven Machine Learning.

Link To Code: https://github.com/olakalisz/adios

Primary Area: Applications->Health / Medicine

Keywords: Opponent Shaping, Antibody Design, Meta Learning, Game Theory, Computational Biology

Submission Number: 12226