EnsembleClust

EnsembleClust performs hierarchical clustering of noncoding RNAs to group unannotated transcripts into putative families based on sequence and secondary-structure similarity.

Key Features:

• Innovative Similarity Measure: Leverages suboptimal solutions within dynamic programming frameworks to define a similarity measure that improves accuracy of approximate ncRNA clustering algorithms.
• Comprehensive Alignment Strategy: Considers all possible sequence alignments and secondary structures rather than a single optimal alignment to assess ncRNA similarity.
• Balanced Performance: Maintains clustering quality while controlling computational cost, enabling performance when sequence identity among family members falls below 60%.
• Fast and Accurate Clustering: Simplifies approximation of structural alignment to achieve faster clustering suitable for large-scale genomic datasets.

Scientific Applications:

• Discovery of novel ncRNA families: Facilitates identification of new noncoding RNA families by grouping unannotated transcripts based on sequence and structure.
• Evolutionary and functional inference: Supports inference of evolutionary relationships and potential functional similarity among ncRNAs with low primary sequence identity.
• Annotation of unannotated transcripts: Assists annotation efforts by clustering unannotated transcripts into candidate families for further analysis.

Methodology:

Uses approximate structural alignment by incorporating suboptimal dynamic-programming solutions and considering all possible sequence alignments and secondary structures, followed by hierarchical clustering.