Fig. 1

The workflow of the RPocket database construction. A 269 RNA-ligand structures were used for analysis. B To acquire the non-redundant dataset, we performed the sequence alignment for the 269 structures using the CD-Hit server. We have used two identity cutoffs (0.80 and 0.95) to get relatively loose and more strict non-redundant datasets: RBL75 and RBL94 (75 and 94 clusters for 0.80 and 0.95 sequence identity cutoffs). C Interaction information and ligands binding sites were identified using Ligplot + and a distance-based calculation. D The functional motifs were identified by the RegRNA program. The RegRNA identifies the RNA motifs by integrating regulatory RNA motifs from the published literature and RNA motif databases. E The pockets were detected by the 3 V server using the rolling probe method. F RNA pocket shape distribution and classification were generated using RPDescriptor. G The ligands functional groups, hydrogen bond and non-bond interactions, the secondary structure patterns, and pocket topology information were calculated and provided in the RPocket server