Protein structure prediction aided by geometrical and probabilistic constraints

Area of Science:

• Computational Biology
• Biophysics
• Structural Bioinformatics

Background:

• Database-assisted ab initio protein structure prediction methods show promise, with successes in community-wide experiments like CASP.
• Accurate protein structure prediction is crucial for understanding biological function and disease.

Purpose of the Study:

• To develop and evaluate a novel computational methodology for ab initio protein structure prediction.
• To employ combinatorial optimization and Monte Carlo minimization techniques to identify stable protein conformations.

Main Methods:

• Utilized a Monte Carlo minimization algorithm on a constrained search space.
• Applied radius of gyration cutoffs, loop backbone dihedral probability distributions, and secondary structure packing conformations for constraint.
• Incorporated geometrical constraints using distance restraint potential functions to simplify the optimization search.

Main Results:

• Generated 1000 initial conformations for several protein sequences.
• Selected the 50 best candidates as probable conformations based on energy minimization.
• Demonstrated simplification of the optimization process through geometrical restraints.

Conclusions:

• The developed methodology offers a simple and modifiable approach to ab initio protein structure prediction.
• The use of combinatorial optimization and specific restraints effectively narrows down the search for optimal protein structures.
• This method holds potential for further refinement by incorporating additional geometric and probabilistic restraints.