Geno-GNN Characterizes Heterogeneous SARS-CoV-2 Fitness Dynamics
Wang et al., published January 12, 2026 in PLoS Computational Biology, introduce Geno-GNN, a graph representation learning model that predicts ACE2 binding affinity and immune escape from SARS-CoV-2 RBD sequences. Applying Geno-GNN to millions of sequences and external deep mutational scanning datasets, they identify two distinct evolutionary fitness trajectories and report that real-world variants predominantly maintain ACE2 affinity while gaining moderate immune evasion.
Key Points
- 1Developed Geno-GNN model accurately predicts ACE2 affinity and immune escape from sequence data
- 2Identified two fitness trajectories: trade-off immune evasion versus maintained ACE2 affinity with moderate escape
- 3Revealed real-world variants predominantly preserve ACE2 affinity, informing surveillance and vaccine design priorities
Scoring Rationale
Peer-reviewed, reproducible model with actionable outputs drives score; novelty limited relative to prior ML genotype–phenotype methods.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
