ProbeAlign

ProbeAlign integrates high-throughput sequencing-based RNA structure probing reactivity data into ncRNA homology searches to improve nucleotide-resolution structural interpretation and thermodynamic folding prediction accuracy.

Key Features:

• High-throughput reactivity integration: Incorporates quantitative nucleotide-resolution reactivity data from high-throughput sequencing-based RNA structure probing.
• Reactivity-guided homology search: Augments ncRNA homology searches with target-sequence probing reactivity to guide alignments.
• Query/socket pattern omission: Ignores specific pairing patterns within query/socket sequences to streamline alignments for large-scale datasets.
• Partial structural guidance: Utilizes partial structural information embedded within target sequences' probing data to inform alignments.
• Alignment transformation: Transforms the structure-alignment problem into a sequence-alignment problem augmented with reactivity information.
• Thermodynamic folding improvement: Improves the accuracy of thermodynamic folding predictions by integrating experimental reactivity constraints.
• Benchmark performance: Demonstrates higher prediction accuracy than filter-based CMsearch methods while maintaining high computational efficiency.
• Large-scale compatibility: Applied to FragSeq genome-wide structure probing datasets and compatible with extensive datasets derived from high-throughput sequencing technologies.

Scientific Applications:

• ncRNA functional annotation: Facilitates functional annotation of non-coding RNAs by incorporating structure probing data into homology searches.
• Genome-wide ncRNA searches: Enables genome-scale ncRNA searches across FragSeq and other high-throughput sequencing-derived structure probing datasets.
• Improved folding predictions: Refines nucleotide-resolution thermodynamic folding predictions using experimental reactivity data.

Methodology:

Incorporates reactivity data from high-throughput RNA structure probing into ncRNA homology searches, ignores specific pairing patterns of query/socket sequences, and uses partial structural information in target probing data to convert structure alignment into a sequence alignment problem augmented with reactivity information.