Learning the language of viral evolution and escape

Abstract

<jats:title>Abstract</jats:title><jats:p>Viral mutation that escapes from human immunity remains a major obstacle to antiviral and vaccine development. While anticipating escape could aid rational therapeutic design, the complex rules governing viral escape are challenging to model. Here, we demonstrate an unprecedented ability to predict viral escape by using machine learning algorithms originally developed to model the complexity of human natural language. Our key conceptual advance is that predicting escape requires identifying mutations that preserve viral fitness, or “grammaticality,” and also induce high antigenic change, or “semantic change.” We develop viral language models for influenza hemagglutinin, HIV Env, and SARS-CoV-2 Spike that we use to construct antigenically meaningful semantic landscapes, perform completely unsupervised prediction of escape mutants, and learn structural escape patterns from sequence alone. More profoundly, we lay a promising conceptual bridge between natural language and viral evolution.</jats:p><jats:sec><jats:title>One sentence summary</jats:title><jats:p>Neural language models of semantic change and grammaticality enable unprecedented prediction of viral escape mutations.</jats:p></jats:sec>

The Prolyl-tRNA Synthetase Inhibitor Halofuginone Inhibits SARS-CoV-2 Infection

“…Convalescent plasma from individuals who have recovered from COVID-19 has also been granted EUA for hospitalized patients with COVID-19 but efficacy data from randomized trials are needed. And despite massive roll-out of vaccines against SARS-CoV-2, the rapid emergence of novel strains amongst other practical issues raises serious concerns on the longevity and efficacy of this approach to confer herd immunity (Hie et al, 2021). Hence, there is an urgent unmet clinical need for an arsenal of potent antiviral therapeutics that can reduce COVID-19 associated morbidity and mortality (McCreary and Angus, 2020; Thorp, 2020).…”

The language of proteins: NLP, machine learning & protein sequences

“…Language models have been used to understand and predict viral mutations that evade neutralizing antibodies [35] . Language generation models can also be applied to synthetic protein design, as in ProGen and other works [56] , [110] , [4] .…”

Overcoming Immunological Challenges Limiting Capsid-Mediated Gene Therapy With Machine Learning

“…If these and other methods can be applied to the problem of AAV capsid engineering, AAV variant sequences with similar properties to each other would be close together in latent space after being transformed into their latent representations, even if they are far apart in sequence space. A similar strategy was recently used to predict the emergence of escape mutations in multiple viruses ( 22 ).…”