PROBE: a computer program employing an integrated neural network approach to protein structure prediction

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein structure prediction

Background:

• Accurate prediction of protein structural features is crucial for understanding protein function.
• Existing methods often predict individual features separately, limiting synergistic improvements.

Purpose of the Study:

• To develop a comprehensive computational tool, PROBE, for predicting diverse protein structural features from amino acid sequences.
• To integrate multiple prediction methods for enhanced accuracy and a holistic view of protein structure.

Main Methods:

• Utilizing a suite of computer-simulated neural networks, each specialized for predicting specific protein structural aspects.
• Implementing a system allowing communication and synergistic comparison between different neural networks within the PROBE program.
• Analyzing amino acid composition and relative hydrophobicity to classify proteins into common folding motifs.

Main Results:

• PROBE predicts surface accessibility, disulfide bonds, and secondary structure composition (including beta-turn types) from amino acid sequences.
• The program can determine a protein's folding motif based on its amino acid composition and hydrophobicity.
• Synergistic interactions between neural networks within PROBE lead to improved prediction accuracy.

Conclusions:

• PROBE offers an integrated and user-friendly platform for predicting multiple protein structural characteristics.
• The program's ability to combine predictions from various neural networks enhances the reliability of protein structure analysis.
• PROBE advances computational approaches for predicting protein structure and folding motifs from sequence data.