SABBAC

SABBAC reconstructs protein backbone structures from alpha-carbon (Cα) trace data by positioning missing backbone atoms while preserving the original Cα coordinates.

Key Features:

• Fragment-Based Assembly: Encodes Cα traces using a hidden Markov model (HMM)-based structural alphabet and selects fragments from a reduced fragment library for reconstruction.
• Greedy Algorithm with Energy Scoring: Assembles selected fragments using a greedy algorithm guided by an energy-based scoring function to favor energetically favorable configurations.
• Robustness and Performance: Demonstrates performance comparable to or better than other protein backbone reconstruction methods and tolerates deviations in Cα coordinates.
• No Further Refinement Required: Produces backbone atom positions that do not require additional refinement after reconstruction.

Scientific Applications:

• Protein Structure Prediction: Reconstruction of complete backbone coordinates from partial Cα trace data to support three-dimensional structure modeling.
• Structural Analysis and Comparison: Generation of full backbone models for comparative structural analysis and studies of conformational relationships.
• Drug Design and Development: Provision of accurate backbone frameworks for modeling target proteins in structure-based drug discovery workflows.

Methodology:

SABBAC encodes the Cα trace into a structural alphabet using HMMs, selects appropriate fragments from a curated reduced library, and assembles them into a complete backbone using a greedy algorithm optimized by energy-based scoring.