Genomic factors shaping codon usage across the Saccharomycotina subphylum.

Area of Science:

• * Genomics and Molecular Biology: Investigating codon usage patterns and their evolutionary drivers in fungi.
• * Bioinformatics and Machine Learning: Applying computational methods to classify yeast orders based on codon bias.

Background:

• * Codon usage bias, the non-uniform selection of synonymous codons, impacts cellular functions like translation and transcript stability.
• * The Saccharomycotina subphylum, including yeasts like *Saccharomyces cerevisiae* and *Candida albicans*, serves as a model for studying these biases.
• * Previous research indicates codon usage is shaped by both translational selection and neutral evolutionary forces.

Approach:

• * Analyzed codon usage across 1,154 strains from 1,051 Saccharomycotina species.
• * Correlated codon usage bias with genomic features such as GC content and tRNA-ome size.
• * Utilized a machine learning algorithm to classify yeast orders based on codon usage patterns.

Key Points:

• * A general preference for A/T-ending codons was observed across the Saccharomycotina.
• * Codon usage bias strongly correlates with GC content and tRNA-ome size.
• * Machine learning accurately classifies yeast orders (>90%) using codon usage data.
• * Translational selection is influenced by genome features like coding sequence count and genome length.
• * The Saccharomycodales order exhibits extreme codon bias due to a lack of arginine tRNAs (CGN codons).

Conclusions:

• * Codon usage bias in Saccharomycotina is shaped by genomic features and GC bias, consistent with prior fungal studies.
• * Specific yeast lineages display extreme codon preferences and avoidance, suggesting unique evolutionary pressures.
• * The lack of arginine tRNAs in Saccharomycodales is associated with the functional decline of CGN codons.