Differentially used codons among essential genes in bacteria identified by machine learning-based analysis

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Codon usage bias (CUB) describes the non-uniform selection of synonymous codons for the same amino acid, varying across bacterial genomes.
• Gene expression levels are known to influence CUB, leading to differences between high- and low-expression genes in bacteria.
• The role of gene essentiality in shaping CUB patterns remains an underexplored area in bacterial genomics.

Purpose of the Study:

• To investigate codon usage patterns in bacteria by incorporating gene essentiality as a key feature.
• To analyze differences in Relative Synonymous Codon Usage (RSCU) between essential and non-essential genes across multiple bacterial species.
• To apply machine learning models to classify genes based on essentiality using CUB features and identify conserved codon usage patterns.

Main Methods:

• Analysis of Relative Synonymous Codon Usage (RSCU) values for essential and non-essential genes.
• Application of machine learning (ML) approaches for gene classification based on codon usage.
• Comparative analysis across 35 bacterial genomes with available gene essentiality data.

Main Results:

• Significant differences in codon usage patterns were observed between essential and non-essential genes in the majority of the studied bacterial genomes.
• Machine learning classifiers achieved high Area Under Curve (AUC) scores (minimum 70.0) for classifying essential genes across 28 organisms.
• Specific codons like CGT (Arg) and GGT (Gly) were highly preferred in essential genes in E. coli, with CGT, ATA, GGT, and GGG showing consistent importance across genomes; TGY (Cys) and CAY (His) codons contributed least.

Conclusions:

• Gene essentiality is a significant factor driving synonymous codon usage differences in bacterial genomes.
• Machine learning effectively distinguishes essential from non-essential genes based on codon usage patterns.
• A common codon usage pattern associated with essential genes exists across diverse bacterial species.