RNAthreader

RNAthreader identifies remote homologies in RNA sequences and facilitates template-based modeling of their three-dimensional (3D) structures by combining RNA secondary structure and sequence profiles.

Key Features:

• Remote homology detection: Identifies remote homologies in RNA sequences to enable template-based 3D structure modeling.
• Secondary structure and sequence-profile integration: Combines RNA secondary structure and sequence profiles to guide threading and alignment.
• Secondary structure simplification: Simplifies complex RNA secondary structures into compact diagrams to reduce computational complexity and memory usage.
• Two-step alignment: Performs an initial alignment using integer programming combined with dynamic programming on the simplified diagram, followed by a refined final alignment using dynamic programming that incorporates the initial alignment.
• Pseudoknot handling: Produces more accurate alignments for RNAs containing pseudoknots compared to conventional methods.
• Benchmarking and accuracy: Validated on 80 RNA sequences with improved alignment accuracy and on 30 RNA-Puzzles cases producing templates that yield lower average root-mean-square deviation (RMSD).
• Computational efficiency: Completes alignments for approximately 5,000 RNAs in under two hours and is reported to be 3–40 times faster than other methods.

Scientific Applications:

• Template-based 3D modeling: Enables selection and alignment of templates for building RNA 3D structure models.
• Remote homology detection in RNAs: Detects distant evolutionary relationships among RNA sequences.
• Modeling RNAs with pseudoknots: Improves alignments and subsequent modeling of RNAs that contain pseudoknots.
• Large-scale RNA template detection: Supports high-throughput identification of templates across thousands of RNAs.

Methodology:

The method simplifies RNA secondary structures into compact diagrams, integrates sequence profiles with secondary-structure information, performs an initial alignment using integer programming combined with dynamic programming on the simplified diagram, and generates a refined final alignment using dynamic programming that incorporates the initial alignment.