LocARNA-P

LocARNA-P computes structure-based multiple alignments and alignment reliability profiles to improve boundary prediction and discrimination of structural noncoding RNAs (ncRNAs).

Key Features:

• Structure-Based Alignment Reliabilities (STARs): Computes columnwise and global reliabilities in RNA multiple alignments based on sequence and structure similarity (STARs).
• Improved Boundary Prediction: Refines boundary predictions from ncRNA gene finders such as RNAz, reducing median boundary deviation from 47 nucleotides to 13 nucleotides.
• Discrimination Between True and False Positives: Uses STAR scores to increase discrimination between true and false ncRNA predictions, raising area under the curve (AUC) from 0.71 to 0.87.
• Comprehensive Analysis Outputs: Produces structure-based multiple RNA alignments accompanied by columnwise STAR profiles to support manual and automated analysis of structural ncRNAs.

Scientific Applications:

• ncRNA boundary refinement: Refining predicted ncRNA boundaries produced by genomic screens and tools like RNAz.
• Candidate evaluation and filtering: Discriminating true structural ncRNAs from false positives in genome-wide ncRNA screens.
• Structural conservation analysis: Characterizing structural conservation in noncoding RNAs through alignment and reliability profiling.

Methodology:

Applies a probabilistic approach to structure-based multiple alignment that integrates sequence and structural similarity to compute columnwise and global alignment reliabilities (STAR profiles).