RNANet

RNANet provides a standardized, integrated dataset combining RNA sequences, homology information, and 3D structural annotations to support machine learning-based RNA structure prediction.

Key Features:

• Integrated Multi-Level Dataset: Combines RNA sequences, position-specific scoring matrices for homologous sequences, and 3D structure annotations including secondary structure, canonical and non-canonical interactions, and backbone torsion angles.
• Automated Parallel Pipeline: Retrieves and integrates data from PDB, Rfam, and SILVA into a reproducible, standardized format with monthly updates.

Scientific Applications:

• Machine Learning for RNA Structure Prediction: Provides curated sequence and structural datasets for training and benchmarking models predicting RNA secondary and 3D structures.

Methodology:

RNANet automatically retrieves RNA sequence and structural data from PDB, Rfam, and SILVA, integrates homologous sequence profiles and detailed 3D annotations into a unified format, and generates statistically described datasets through a parallelized processing pipeline.