Computing the geometry of a molecule in dihedral angle space using n.m.r.-derived constraints. A new algorithm based...

Area of Science:

• Structural biology
• Biophysics
• Computational chemistry

Background:

• Nuclear Magnetic Resonance (NMR) spectroscopy is crucial for determining protein structures.
• Current methods often rely on Cartesian coordinates, which can be computationally intensive.
• Accurate error estimation is vital for validating protein structural models.

Purpose of the Study:

• To develop an efficient method for protein 3D structure determination using NMR constraints.
• To implement a novel approach utilizing dihedral angles for molecular representation.
• To directly provide error estimates on refined structural parameters.

Main Methods:

• Developed a method based on optimal filtering and dihedral angle internal representation.
• Encoded the algorithm in a software package named FILMAN.
• Applied FILMAN to protein folding and peptide cyclization problems.

Main Results:

• Dihedral angle representation significantly reduces computational burden compared to Cartesian coordinates.
• FILMAN allows for better incorporation of a priori information, including proline residue constraints and vicinal coupling constants.
• Demonstrated performance on a 10-residue alanine polypeptide, an 11-residue peptide cyclization, and the medium-sized protein tendamistat.

Conclusions:

• FILMAN provides an efficient and robust method for protein structure determination from NMR data.
• The direct error estimation capability enhances the reliability and interpretability of the determined structures.
• The use of dihedral angles offers advantages in computational efficiency and control over input constraints.