The trRosettaRNA server for RNA structure prediction

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• RNA molecules fold into complex three-dimensional (3D) structures essential for their biological functions, analogous to proteins.
• Accurate prediction of these RNA 3D structures is crucial for understanding their roles in cellular processes.

Purpose of the Study:

• To introduce trRosettaRNA, a novel web-based platform for automated RNA 3D structure prediction utilizing deep learning.
• To provide researchers with a tool for accurate and efficient prediction of RNA tertiary structures.

Main Methods:

• Development of an end-to-end deep learning neural network for direct RNA 3D structure prediction from nucleotide sequences.
• Integration of optional inputs for custom multiple sequence alignments and secondary structures to enhance prediction accuracy.
• Inclusion of an energy optimization step to refine predicted structures and resolve potential violations.

Main Results:

• The trRosettaRNA server demonstrates state-of-the-art modeling accuracy for RNA 3D structure prediction.
• Predictions for RNA families lacking known 3D structures showed high confidence and good agreement with experimental data.
• Median prediction time for a 200-nucleotide RNA sequence is approximately 1 hour using 5 CPU cores.

Conclusions:

• trRosettaRNA offers a powerful, accurate, and user-friendly solution for automated RNA 3D structure prediction.
• The availability of both a web server and an open-source standalone package enhances accessibility, data privacy, and customization for diverse research needs.
• trRosettaRNA facilitates advancements in understanding RNA structure-function relationships and enables the development of new bioinformatics tools.