Neural network in the transcription-translation process

Area of Science:

• Computational Biology
• Molecular Biology
• Bioinformatics

Background:

• The genetic code's degeneracy, where multiple codons can specify the same amino acid, is a fundamental aspect of molecular biology.
• Understanding the relationship between genetic code properties and protein characteristics is crucial for deciphering evolutionary constraints and functional implications.

Purpose of the Study:

• To model the biological process of transcription and translation using a virtual neural network.
• To investigate the correlation between natural amino acid frequency, codon degeneracy, and protein characteristics.
• To explore the impact of genetic code degeneracy on protein evolution and stability.

Main Methods:

• Development of a virtual neural network to represent the transcription-translation process.
• Input layer representing transcription, hidden layer for mRNA-ribosome-tRNA interactions, and output layer for the final protein state.
• Analysis of correlations between amino acid frequencies, codon degeneracy, and protein substitution rates.

Main Results:

• A direct correlation was identified between the natural frequency of amino acids and the degree of codon degeneracy.
• An inverse correlation was observed between code degeneracy and the number of amino acid substitutions within a protein.
• The neural network model successfully simulated key aspects of protein synthesis.

Conclusions:

• The study provides a computational model linking genetic code properties to protein synthesis and evolution.
• Codon degeneracy appears to play a significant role in modulating protein stability and evolutionary pathways.
• Findings highlight the intricate relationship between the genetic code's structure and the resulting proteome.