Characterization of the heterogeneity in SARS-CoV-2 fitness dynamics via graph representation learning

TL;DR

The paper introduces Geno-GNN, a deep learning tool that helps understand how SARS-CoV-2 adapts to immunity by predicting its ability to infect cells and evade the immune system.

Contribution

Geno-GNN is a novel graph-based deep learning framework that accurately predicts ACE2 binding and immune escape of SARS-CoV-2.

Findings

SARS-CoV-2 variants predominantly maintain ACE2 binding while achieving moderate immune evasion.

Two distinct fitness trajectories were identified: immune evasion at the cost of infectivity or balanced evasion with maintained infectivity.

Geno-GNN reveals complex evolutionary patterns of SARS-CoV-2 driven by population immunity contexts.

Abstract

Understanding the heterogeneity of population-level viral fitness dynamics, which reflect the interplay between intrinsic viral properties and population immunity, is critical for pandemic preparedness. However, how these dynamics vary across diverse immune backgrounds and mutational landscapes remain poorly characterized. We present Geno-GNN, a graph representation learning approach for retrospectively characterizing the viral fitness dynamics of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Geno-GNN accurately predicts angiotensin-converting enzyme 2 (ACE2) binding affinity and immune escape potential across multiple external datasets. Using Geno-GNN, we identified temporal patterns in SARS-CoV-2 fitness and detected varying rates of fitness change associated with distinct immune backgrounds. Virtual mutation scanning revealed two fitness trajectories: broad immune evasion at the cost of ACE2 affinity and ACE2 affinity maintenance at or above the Wuhan-Hu-1 level along with moderate immune escape. Notably, real-world SARS-CoV-2 variants predominantly followed the latter trajectory, sustaining ACE2 affinity via fixed mutations. These findings underscore the heterogeneous, immune-contextualized nature of viral fitness dynamics and the complex evolutionary pathways of SARS-CoV-2.

Understanding how viruses adapt to population immunity is essential for managing pandemics. Here, we developed Geno-GNN, a deep learning framework that accurately predicts two critical viral fitness from sequence data: the ability to bind to human cells and the capacity to evade immunity. Applying Geno-GNN to SARS-CoV-2, we found that the rate of change in its fitness varied in response to different contexts of population immunity. Two fitness strategies were identified: sacrificing infectivity for stronger immune escape, or maintaining infectivity while achieving moderate immune evasion. Real-world SARS-CoV-2 variants have followed the latter, balanced trajectory, maintained by the synergistic effect of multiple mutations acting in concert. By capturing these complex interaction, Geno-GNN provides a useful tool to systematically explore viral sequence space and to support preparedness for future emerging pathogens.

Figures