Fig. 3: Comparison of the prediction accuracy on the SCOPe v2.08 S95 structural classification database.
From: An end-to-end framework for the prediction of protein structure and fitness from single sequence
a, b In the scatter plot, each point represents the average TM-score of all domains of one fold in SCOPe. The vertical axis is the prediction performance (TM-score) of SPIRED on different SCOPe folds, while the horizontal axis represents results for OmegaFold and ESMFold, respectively. c The KDE plot is used to visualize the distribution of model performance (in TM-score) over different SCOPe folds for the tested models. d The boxplot compares the TM-scores calculated over all SCOPe folds (sample size = 1231) for different models, where the green triangle indicates the mean value with the corresponding number marked aside, the center line represents the median, the box limits correspond to the upper and lower quartiles and the whiskers extend to 1.5 inter-quartile range. Source data are provided as a Source Data file.
