RNASSTR v2: RNA sequence-secondary-structure pairs and model benchmark outputs

RNASSTR v2 provides structure-aware training, validation, and test partitions of RNA sequence-secondary-structure pairs. The release also includes the prediction and scoring files used to evaluate published SincFold, RNASSTR-retrained SincFold, Lyra-TransPred, and RNAfold in the accompanying revised manuscript, together with structural-split provenance and metadata tables. Dataset-generation code and documentation are available at https://github.com/romanagle/RNASSTR. Langeberg, C. J. et al. Improving RNA Secondary Structure Prediction Through Expanded Training Data. bioRxiv (2025). https://doi.org/10.1101/2025.05.03.652028