Fig. 1

Degron collection, processing, and model performance. a Degrons were collected from the true positive degrons in the ELM database and three previous studies. Venn diagram showing the overlap of collected degrons from three previous studies. b Data augmentation: sampling 128AA-peptides around degrons from original protein sequences. The blue rectangles represent degrons, and the black lines represent peptide sequences. c Training sets, test sets, and performances of models, models were evaluated with the area under the receiver operating characteristic curve (AUC). d Degron clustering and composition of five clusters, Seqlogos in the upper panel show the sequence patterns of alignment cores for five clusters in Gibbscluster, classes accounting for at least 10 percent in each cluster are labeled in pie graphs