Incorporating Ab Initio energy into threading approaches for protein structure prediction

Area of Science:

• Computational Biology
• Structural Bioinformatics

Background:

• Protein structure is determined by local and distant residue interactions.
• Threading and Ab Initio methods predict protein structures using interaction information.
• Structure refinement benefits from potentials capturing distant interactions.

Purpose of the Study:

• Investigate using Ab Initio energy functions to improve protein threading alignment.
• Enhance threading for weakly or distantly homologous templates.

Main Methods:

• Incorporate distant interaction information into threading alignment scoring.
• Utilize an Ab Initio energy function to evaluate partial decoys from alignments.
• Employ a local search algorithm for scoring function optimization.

Main Results:

• Improved alignment quality for 68 out of 127 query-template pairs in the Prosup benchmark.
• Outperformed state-of-the-art threading methods in alignment accuracy.

Conclusions:

• Integrating Ab Initio energy functions significantly enhances protein threading alignment accuracy.