ThreaDomEx

ThreaDomEx predicts continuous and discontinuous protein domain structures by threading sequences through the Protein Data Bank (PDB) and applying domain conservation scoring and assembly algorithms for accurate domain boundary delineation.

Key Features:

• Hierarchical Pipeline Approach: Threads query protein sequences through the Protein Data Bank (PDB) to identify structural templates as the basis for downstream domain predictions.
• Domain Conservation Score (DC-score): Derives a profile of DC-scores during template identification to assign domain segments and delineate domain boundaries.
• Boundary Clustering Algorithm: Uses a boundary-clustering algorithm to detect and refine discontinuous domains (DDCs) by improving DCD-linker location predictions.
• Symmetry-Guided Domain-Segment Assembly: Applies symmetry comparison-guided assembly to detect and integrate DCDs when templates lack DCDs.

Scientific Applications:

• Benchmarking on 1,111 proteins: Evaluated on a dataset of 1,111 proteins, achieving a normalized domain overlap score of 89.3% versus experimental data.
• DCD detection performance: Recalls 26.7% of DCDs with 72.7% precision on proteins where initial threading failed to detect any DCDs.
• Comparative performance: Demonstrates higher normalized domain overlap than other state-of-the-art methods in the reported benchmark.

Methodology:

Integrates threading alignments, DC-score profiling, boundary clustering, and symmetry-guided assembly.