UniDoc

UniDoc parses multi-domain proteins into individual domains by integrating sequence and structure information to delineate domain boundaries for structural and functional studies.

Key Features:

• Unified sequence- and structure-based parsing: integrates sequence-based and structure-based domain parsing within a single framework.
• Inter-residue distance matrix: leverages inter-residue distance matrices sourced from experimental structures or predicted by trRosettaX.
• Deep learning-based inputs: uses outputs from deep learning-based protein structure prediction methods such as trRosettaX.
• Physical optimization principle: defines domains by maximizing intra-domain interactions while minimizing inter-domain interactions.
• Applicability: supports both sequence-based and structure-based domain parsing approaches.
• Benchmark performance: demonstrated higher accuracy and speed than other state-of-the-art methods on five benchmark datasets.

Scientific Applications:

• Domain boundary delineation: precise parsing of multi-domain proteins for structural studies.
• Functional domain analysis: supports domain-based functional annotation and analysis.
• Structural biology research: application in structural biology and related studies that require domain-level decomposition.

Methodology:

UniDoc uses inter-residue distance matrices from experimental structures or predicted by trRosettaX, integrates sequence and structure information, and applies the principle of maximizing intra-domain interactions while minimizing inter-domain interactions to define domain boundaries.