COMER

COMER detects protein remote homology using profile-profile alignment with an enhanced substitution scoring scheme to improve sensitivity and alignment quality for comparative modeling and protein structure prediction.

Key Features:

• Profile-Profile Alignment: Accepts a multiple sequence alignment (MSA), converts it to a profile representation, and searches against a database of profiles to identify statistically significant similarities.
• Refined Substitution Scores: Implements an innovative add-on score that augments traditional profile-profile substitution scores to increase alignment sensitivity and quality.
• Statistical Improvement: Comprehensive evaluations demonstrate that the add-on score yields statistically significant improvements in alignment outcomes.
• Computational Complexity: The scoring method is designed with almost optimal computational complexity to enable integration into existing profile-profile alignment methods without compromising performance.

Scientific Applications:

• Comparative Modeling: Facilitates comparative modeling for protein structure prediction by improving remote homology detection and alignment quality.
• Structure-Function Inference: Supports prediction of protein structures from known sequences and downstream inference of protein functions and interactions.

Methodology:

Transforms an MSA into a profile, applies a refined profile-profile substitution scoring mechanism (an innovative add-on score), searches a database of profiles to identify statistically significant similarities, and evaluates alignment improvements while maintaining almost optimal computational complexity.