CMCompare

CMCompare compares covariance models (CMs) produced with Infernal to evaluate specificity and interrelationships among RNA family models for non-coding RNA (ncRNA) identification.

Key Features:

• Model Specificity Assessment: Identifies covariance models with poor specificity by detecting cross-hits and non-homologous matches when compared to reference models such as those in Rfam.
• Exploration of Model Relationships: Explores similarities, overlaps, and relationships between two or more CMs to reveal potential family-level connections and redundancies.
• Avoidance of Duplicate Models: Compares candidate CMs against existing Rfam models to detect redundant models representing the same ncRNA family.
• Clan Evaluation Support: Reveals family-to-family relationships to support identification and evaluation of clans (clusters of biologically related families).

Scientific Applications:

• Novel ncRNA identification: Supports identification and characterization of novel non-coding RNAs by validating and refining family covariance models.
• Homology searches: Improves the quality of homology searches by ensuring model specificity and uniqueness.
• Classification and clustering: Aids classification and clustering of ncRNAs into biologically meaningful groups through analysis of model relationships.

Methodology:

CMCompare compares covariance models, including those constructed with Infernal, by evaluating shared sequence and structural features against a database of existing models such as Rfam (currently including 2,208 families).