Fig. 5
From: Labeled dataset of X-ray protein ligand images in 3D point cloud and validated deep learning models
Class distribution of SP-based vocabularies. Distribution of the class occurrence in the SP-based vocabularies by labeled atom of all entries of the final list of valid ligands. Their corresponding imbalance ratio (dmax) is also presented. The distribution for the “Vocabulary of the Ligand Region” is omitted because all 2,566,614 atoms were labeled with the same generic class of atoms. The background class is not used in these distributions.
