PETra
A Transformer model for predicting future SARS-CoV-2 mutations, described in a sole-author preprint on arXiv (2025).
PETra (Pretrained Evolutionary Transformer) is a deep learning model for predicting future mutations in SARS-CoV-2. Pretrained on large-scale viral genomic data, it frames mutation prediction as a sequence modeling problem: given the evolutionary history of the virus, which amino acid substitutions are most likely to emerge next?
Unlike traditional tree-based phylogenetic inference, PETra learns evolutionary patterns directly from sequence data.
Try the online tool at cpredict.cn; method details are in the arXiv preprint.