Energy Profile Bayes and Thompson Optimized Convolutional Neural Network protein structure prediction.

Area of Science:

• Computational Biology
• Bioinformatics
• Structural Bioinformatics

Background:

• Proteins execute biological functions, making protein structure comprehension critical.
• Experimental methods are insufficient for predicting protein structural classes due to data volume and complexity.
• Computational methods are advantageous for predicting protein structural classes from sequences.

Purpose of the Study:

• To propose a reliable computational method for predicting protein structural classes from protein sequences.
• To leverage deep learning for enhanced protein structure prediction.
• To introduce the Energy Profile Legion-Class Bayes Protein Structure Identification model.

Main Methods:

• Developed the Energy Profile Legion-Class Bayes Protein Structure Identification model.
• Employed a Thompson Optimized convolutional neural network for amino acid feature extraction.
• Utilized the Thompson Optimized SoftMax function for protein sequence association analysis.
• Proposed the Energy Profile Bayes and Thompson Optimized Convolutional Neural Network (EPB-OCNN) method.

Main Results:

• The EPB-OCNN method was tested on diverse protein data.
• Performance was compared against state-of-the-art methods including Template-Based Modeling and deep learning approaches.
• Evaluated metrics included prediction time, accuracy, specificity, recall, F-measure, and precision using the Biopython tool.

Conclusions:

• The EPB-OCNN method demonstrated superior performance compared to existing state-of-the-art methods.
• The proposed method effectively predicts protein structural classes, addressing limitations of experimental techniques.
• Deep learning integration significantly enhances protein structure prediction capabilities.