EXTRARNAS: Data, Results and Analysis Scripts

EXTRARNAS – Data, Results and Analysis Scripts Companion material for the CIBB 2026 short paper This repository contains the datasets, execution outputs, analysis scripts, and supplementary material accompanying the paper EXTRARNAS: A Framework for Extracting RNA Structures with Multiple Tools The EXTRARNAS software itself is distributed separately. Software repository: https://github.com/bdslab/EXTRARNAS Software release (version used in the paper): https://doi.org/10.5281/zenodo.21238912 Repository structure · experiment_triple_helices/ o shared_folder_before_execution/ o shared_folder_after_execution/ o script_input_folder/ o script_output_folder/ o adb-reference/ o bpseq-reference/ o extrarnas_compare_triple_helix_or_general_with_precision_recall.py · additional-experiments/ o 5s-pseudoknot-batch/ · README.md Contents experiment_triple_helices Material used for the evaluation reported in the paper. shared_folder_before_execution Initial state of the shared Docker folder before executing EXTRARNAS. Contains: · input CSV; · local PDB structures not retrieved automatically from the PDB. shared_folder_after_execution Complete output generated by EXTRARNAS after processing the eight RNA triple-helix structures. Includes the raw outputs produced by all supported annotation tools. script_input_folder Input used by the analysis script. Contains: · BPSEQ files; · BPSEQE files; · curated BPSEQ references; · augmented dot-bracket (ADB) references. script_output_folder Output generated by the comparison script. Includes: · summary_by_molecule.csv · aggregate_by_tool.csv · pairwise_tool_overlap.csv · missing_extra_pairs.tsv · aggregate_table.tex · paper_table.tex · supplementary.tex · supplementary.pdf The supplementary PDF contains the complete per-molecule evaluation table used to generate the aggregate results reported in the paper. Reference datasets The folders · bpseq-reference · adb-reference contain the manually curated reference annotations introduced in Matarrese et al., Decoding RNA Triple Helices (2026). additional-experiments Contains additional experiments not included in the quantitative evaluation of the paper. 5s-pseudoknot-batch Batch-processing experiment on 20 RNA 5S structures. This experiment documents the execution of EXTRARNAS on an independent dataset and includes: · input dataset; · execution log; · preprocessing results; · outputs generated by each annotation tool. Some structures could not be fully processed because of current limitations of the mmCIF→PDB preprocessing pipeline (BeEM bundled conversion), rather than failures of the EXTRARNAS parsing or comparison workflow. Requirements The comparison scripts require only: · Python 3.8 or newer; · standard Python libraries. Docker is required only to reproduce the complete EXTRARNAS workflow. Running the comparison script From the experiment_triple_helices directory: python3 extrarnas_compare_triple_helix_or_general_with_precision_recall.py \--input-dir script_input_folder \--output-dir script_output_folder By default, the script analyzes the eight RNA molecules used in the paper. Use --all to analyze every available structure or --molecule <name> to analyze a single RNA molecule. Reproducibility The repository contains all datasets, reference annotations, execution outputs, and analysis scripts required to reproduce the evaluation reported in the accompanying paper.