aCMs

aCMs implements ambivalent covariance models to perform homology search for structured RNAs by representing multiple consensus secondary structures within a single model to improve detection of non-coding RNA families.

Key Features:

• Support for Multiple Consensus Structures: Represents several compatible consensus secondary structures in a single covariance model to capture structural diversity within RNA families.
• Enhanced Sensitivity and Specificity: Integrates structural variation into the model architecture to improve detection sensitivity while maintaining specificity, enabling more comprehensive representation in databases like RFAM.
• Reduction of Artificial Subdivision: Coalesces structural variants within one model to avoid artificial subdivision of RNA families that occurs when using classical covariance models.

Scientific Applications:

• Non-coding RNA family homology search: Improves identification and characterization of non-coding RNA families by accommodating alternative consensus secondary structures during homology search.
• tRNA structural variation analysis: Captures variations such as additional helical structures beyond the typical cloverleaf configuration in transfer RNAs (tRNAs), enabling more accurate evolutionary and functional analysis.

Methodology:

Extends the fundamental architecture of covariance models to incorporate multiple consensus secondary structures and evaluates model performance on several RFAM families, comparing results to traditional covariance models that often require subdividing families.