COMA

COMA performs sequence profile–profile comparisons to detect distant homologs among protein sequences for inferring their structural and functional relationships.

Key Features:

• Profile-Profile Comparison: Represents protein families as sequence profiles and compares profiles to identify remote homologs.
• Position-Dependent Gap Penalties: Implements position-dependent gap penalties during profile alignment to improve alignment accuracy.
• Global Score System: Incorporates a global score system that provides an analytical solution for estimating statistical parameters and the significance of profile-profile similarities.
• Database Targets: Performs comparisons against structural and family databases including SCOP, PDB, and PFAM to detect homologous relationships.
• Benchmarked Evaluation: Evaluated against HHsearch, COMPASS, and PSI-BLAST using a challenging SCOP domain dataset with at most 20% sequence identity within a reference-free model-based framework.

Scientific Applications:

• Homology Detection: Detects distant homologs to aid inference of protein evolutionary relationships, structure, and function.
• Annotation and Homology Modeling: Supports annotation of uncharacterized proteins and identification of templates for homology modeling.
• Protein Superfamily Identification: Facilitates identification of new members of protein superfamilies, exemplified by applications to the PD-(D/E)XK nuclease superfamily.

Methodology:

Uses sequence profile–profile comparison with sequence profiles representing protein families, position-dependent gap penalties during profile alignment, a global score system with an analytical solution for statistical parameters, and benchmarking within a reference-free model-based framework against HHsearch, COMPASS, and PSI-BLAST using SCOP domains (≤20% sequence identity); comparisons include searches against SCOP, PDB, and PFAM.