Predicting protein variants with equivariant graph neural networks

Antonia Boca, Simon Mathis

Pre-trained models have been successful in many protein engineering tasks. Most notably, sequence-based models have achieved state-of-the-art performance on protein fitness predict…