CCFold

CCFold predicts coiled-coil structures from protein sequences using a threading-based algorithm to model ~300-residue α-helical dimers characteristic of intermediate filaments (IFs).

Key Features:

• Threading-based algorithm: Threads input sequences onto structural fragments to generate coiled-coil models.
• Statistical potentials: Employs statistical analysis derived from experimentally determined coiled-coil structures to guide model building.
• Sequence-to-structure modeling: Generates structural models directly from primary amino acid sequences.
• Long coiled-coil handling: Designed to model long (~300-residue) α-helical coiled-coil dimers typical of IF proteins.
• Hydrophobic repeat flexibility: Accommodates a variety of hydrophobic repeat patterns beyond canonical heptads.
• Performance: Reports higher accuracy and orders-of-magnitude faster runtime compared with general-purpose folding methods.
• Rosetta integration: Integrates with Rosetta folding techniques for further model refinement.

Scientific Applications:

• IF dimer structure prediction: Produces representative dimer models across all classes of intermediate filament proteins.
• Filament assembly studies: Facilitates analysis of structural features relevant to IF assembly mechanisms.
• Mechanical property investigations: Provides models to explore structural determinants of IF mechanical behavior.
• Disease mutation analysis: Enables assessment of the structural implications of disease-related mutations in coiled-coil regions.

Methodology:

CCFold uses a threading-based algorithm guided by statistical analysis of experimentally determined coiled-coil structures to generate sequence-derived models, with integration of Rosetta folding techniques for additional refinement.