bpRNA-CosMoS

bpRNA-CosMoS uses k-mer based cosine similarity to compute similarity scores between RNA secondary structures, enabling fast and robust identification of structural similarities across large datasets.

Key Features:

• bpRNA structure array foundation: The method represents the structural category of each nucleotide as a single-character structural code (e.g. hairpin=H) within the bpRNA structure array framework.
• Cosine similarity of k-mer count vectors: Structural comparison scores are calculated through cosine similarity of k-mer count vectors generated from structure arrays.
• Length-weighted penalty: A length-weighted penalty addresses cases of two RNAs of vastly different lengths, overcoming a major challenge where k-mer based methods often ignore sequence length.
• Fuzzy counting: Optional fuzzy counting adds flexibility to decrease the negative impact that small structural variations have on the similarity score.

Scientific Applications:

• RNA structural comparison across large datasets: The method enables efficient identification of structural similarities in large-scale RNA structural data.
• Alternative to sequence-based comparisons: bpRNA-CosMoS provides a structural similarity approach as an alternative to nucleotide sequence alignments for connecting RNA structure to function.

Methodology:

bpRNA-CosMoS uses k-mers and length-weighted cosine similarity to compute similarity scores between RNA structures represented as bpRNA structure arrays with single-character structural codes. Structural comparison scores are derived from cosine similarity of k-mer count vectors, with optional fuzzy counting to accommodate small structural variations.