LaRA 2

LaRA 2 performs sequence-structure alignment of RNA sequences to integrate secondary-structure information, including pseudoknots and Watson-Crick interactions, for comparative analysis and functional inference of non-coding RNAs.

Key Features:

• Structural alignment: Incorporates RNA secondary structure and Watson-Crick interactions into alignment scoring to align sequences by both sequence and structure.
• Handling pseudoknots: Supports arbitrary pseudoknots to account for overlapping base-pair interactions in secondary structures.
• Performance optimization: Re-implementation employs multi-threading and vectorization for parallel execution, achieving up to 130× speedup over the previous version.
• Algorithmic improvements: Introduces a novel heuristic to compute a lower boundary of the solution, improving performance and accuracy in pairwise and multiple sequence-structure alignments.
• SeqAn implementation: Built on the SeqAn framework to leverage efficient sequence-analysis data structures and algorithms.

Scientific Applications:

• Large-scale RNA comparative analysis: Enables analysis of large datasets of RNA sequences using structure-aware alignments.
• Structural motif discovery: Identifies conserved secondary-structure motifs within RNA molecules.
• Genomic database searches: Uses detected motifs as markers for searching genomic databases for homologous RNAs.
• Functional inference and annotation: Facilitates exploration and annotation of RNA functions, including non-coding RNAs, across biological contexts.

Methodology:

Implemented on the SeqAn framework and uses multi-threading, vectorization, and a novel lower-bound heuristic to compute pairwise and multiple sequence-structure alignments that include arbitrary pseudoknots.