Rnall

Rnall predicts local RNA secondary structures (LSSs) across genomic-scale sequences using a sliding-window dynamic programming approach to enable mining of RNA structural motifs.

Key Features:

• Dynamic Programming Technique: Implements dynamic programming using nearest-neighbor thermodynamic parameters to predict RNA secondary structures.
• Sliding Window Approach: Scans RNA sequences with a sliding window to extract all LSSs no larger than the specified window size.
• Computational Efficiency: Exhibits worst-case time complexity O(W^3L) and observed practical complexity O(W^2L), where W is window size and L is sequence length.
• Energy Landscape Concept: Introduces an energy landscape representation to illustrate local structure stability and support motif mining.
• High Prediction Accuracy: Demonstrates superior prediction accuracy compared to Lfold and Quickfold.

Scientific Applications:

• Genomic-scale RNA structural motif mining: Enables mining of RNA structural motifs across whole genomes by predicting local secondary structures.
• Motif identification: Aids identification of biologically important motifs by producing predicted LSSs that align with known RNA motifs.
• Functional inference of RNAs: Supports studies of RNA molecule function through localized secondary structure prediction.

Methodology:

Uses dynamic programming with nearest-neighbor thermodynamic parameters and a sliding-window scan to extract all LSSs within the window, introduces an energy landscape concept, and reports worst-case complexity O(W^3L) with observed practical complexity O(W^2L).