Fig. 2
From: A prediction of mutations in infectious viruses using artificial intelligence
Phylogenetic tree based on clades, mutation information, and model training data composition. A Nextstrain data collected from December 23, 2019, to February 27, 2024, is visualized by clade, WHO-designated name, and Pango nomenclature lineage. Mutation information found with high frequency (80%) at 72 positions within the receptor binding motif (RBM) of the spike protein is also displayed. Training sets and validation sets are constructed for each clade, forming six stages of training sets. B For the training set and validation set, the mutation rates at each RBM position are displayed