DisCovER

DisCovER performs covariational threading of query protein sequences onto weakly homologous structural templates by integrating inter-residue distance and orientation information and topological network neighborhood effects.

Key Features:

• Distance- and Orientation-Based Threading: Incorporates inter-residue distances and orientations into threading to capture spatial relationships beyond pairwise proximity.
• Topological Network Neighborhood Integration: Includes topological network similarity terms to account for neighborhood effects induced by query-template alignments.
• Iterative Double Dynamic Programming Framework: Employs an iterative double dynamic programming framework to perform and refine distance- and orientation-based alignments.
• Profile-Based Template Selection with Topological Similarity: Selects a subset of candidate templates using profile-based threading augmented by topological network similarity terms.

Scientific Applications:

• Structural Bioinformatics and Protein Modeling: Applied to modeling protein structures with weak sequence homology by improving template selection and alignment accuracy.
• Benchmarking on Hard Targets and CAMEO: Used for benchmarking on hard targets, including the Continuous Automated Model Evaluation (CAMEO) experiment, demonstrating performance improvements over several state-of-the-art threading methods.

Methodology:

Template candidates are selected using profile-based threading augmented with topological network similarity terms, and alignments are refined using an iterative double dynamic programming framework that incorporates inter-residue distance and orientation data.