A test of lattice protein folding algorithms

Area of Science:

• Computational biology
• Protein folding
• Biophysics

Background:

• Protein structure prediction is a fundamental challenge in biology.
• Lattice models offer simplified frameworks for studying protein folding dynamics.
• Designing protein sequences to fold into specific structures is a complex inverse problem.

Purpose of the Study:

• To evaluate lattice-model-based search strategies for identifying global minima in protein chains.
• To test the efficacy of "hydrophobic zippers" and constraint-based hydrophobic core construction (CHCC) methods.
• To assess the success of protein design strategies based on the HP model.

Main Methods:

• Generating target conformations for 48-mer protein chains on a 3D cubic lattice.
• Designing HP (hydrophobic-polar) sequences using an inverse folding algorithm.
• Applying "hydrophobic zippers" and CHCC search methods to find global minima.
• Comparing found conformations with designed target structures.

Main Results:

• The CHCC method successfully found global minima for all tested sequences.
• The "hydrophobic zippers" method found global minima in some cases.
• In 90% of cases, CHCC identified conformations with lower energy than designed.
• Global energy minima exhibited high degeneracy (10^3 to 10^6 conformations).

Conclusions:

• Lattice-model-based search strategies, particularly CHCC, are effective for finding global minima in simplified protein models.
• Current protein design strategies may not perfectly align with the identified global energy minima.
• The study provides insights into conformational space searching and the relationship between simplified models and real proteins.