Prediction of loop regions in protein sequence

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Protein structure prediction is crucial for understanding protein function.
• Identifying regular secondary structures and nonregular loop regions is a key challenge.
• Loop regions play significant roles in protein-protein interactions and antigenicity.

Purpose of the Study:

• To develop and validate an algorithm for decomposing protein chains into regular and nonregular (loop) regions.
• To define optimal parameters for loop prediction using physical properties and entropy scales.
• To evaluate the algorithm's performance against existing methods.

Main Methods:

• An algorithm was developed to input protein sequences and output structural decomposition.
• Analysis of 3,769 globular protein domains to determine optimal prediction parameters.
• Parameters optimized include thresholds for regular/nonregular regions and window sizes for averaging.
• Method utilizes scales of expected contacts and entropy for amino acid flexibility (phi, psi, chi angles).

Main Results:

• The algorithm decomposes protein chains into regular (secondary structure) and nonregular (loop) regions.
• Performance comparison shows similar results to the ALB method (64% true predictions vs. 66%).
• Prediction accuracy for regular and nonregular regions is lower than PSIPRED (73% true predictions).

Conclusions:

• The proposed algorithm offers a novel approach to protein chain decomposition.
• Predicted loop regions can be analyzed for patterns related to structure, function, and antigenicity.
• Further refinement may improve prediction accuracy compared to established methods.