Exploring the Power and Challenges of Predicting Secondary Structures in Ligand-bound RNAs

Data for the manuscript: Exploring the Power and Challenges of Predicting Secondary Structures in Ligand-bound RNAs The dataset was constructed for benchmarking RNA secondary structure predictions on ligand-bound RNAs. All entries have experimentally resolved 3D structures. The data of the self-curated ligand-bound RNA dataset: - Sequences and secondary structures (dot-bracket format), "seq_sec_x3dna.tar.gz" and "seq_sec_fr3d.tar.gz" as the compressed files (annotated by X3DNA and FR3D, respectively), "*.ss" file for each ligand-RNA complex. The first line of each ".ss" file contains the PDB ID and sequence length. The second one is a sequence The third line is secondary structure in dot-bracket format. - Ligand-contacted nucleotides, "ligand_contacted_nt.tar.gz" as the compressed file, "*.lcnt" file for each entry. The residue IDs of ligand-contacted nucleotides are separated by ",". - Non-canonical base pairs, "noncanonical_BPs_x3dna.tar.gz" and "noncanonical_BPs_fr3d.tar.gz" as the compressed files, "*.ncbps" file for each entry. Each row contains the residue IDs of the two nucleotides forming a non-canonical base pair. - WC-edge ligand-contacted bases for each entry, "wc_contact_res.dat" file. Each row contains one PDB ID and its WC-edge ligand-contacted base IDs. If no WC-edge ligand-contacted base exists in this entry, the second column is left empty.