Codon Usage Bias

Research question

How does synonymous codon usage vary across Escherichia coli coding sequences, and how much of that variation can nucleotide composition explain?

Scientific background

Approximately 250 Escherichia coli genome assemblies were obtained from NCBI RefSeq and analyzed computationally. Synonymous codons encode the same amino acid but need not occur at equal frequencies; this project examined those preferences and their possible evolutionary influences.

Methods

The completed undergraduate analysis used sequence preprocessing and descriptive and statistical comparisons to investigate synonymous codon preferences.

• Genome acquisition and preprocessing from NCBI RefSeq
• Protein-coding sequence extraction, filtering, and in-frame codon counting
• Relative Synonymous Codon Usage (RSCU)
• Chi-square goodness-of-fit tests and standardized residual analysis
• Effective Number of Codons (ENC) and GC content at third codon positions (GC3)
• ENC–GC3 comparison with a composition-based expectation
• Statistical analysis and visualization in Python and R

Computational tools

• Python
• R
• Biopython
• tidyverse
• seqinr
• ggplot2
• Linux / Bash

Research progress & findings

The analysis identified nonrandom patterns of synonymous codon usage. ENC–GC3 analysis suggested that nucleotide composition alone may not fully explain the observed bias; estimator and filtering issues in the retained plot require review before stronger inferences.

Figures & visualizations

FIG. 01 · Relative Synonymous Codon Usage (RSCU)

FIG. 02 · Codon bias residuals

FIG. 03 · ENC–GC3 analysis

Interpretation

Selective influences are possible, but these analyses alone do not establish translational selection as the dominant cause. The official poster title is retained in the research-output record; the interpretation here is deliberately more cautious.

Limitations

Results depend on annotation quality, sequence filtering, codon-count assumptions, and the statistical null model. The retained ENC plot includes estimates outside the conventional range. Composition-based comparisons and deviations from equal codon usage cannot, by themselves, distinguish all mutational and selective mechanisms.

Current status

Completed undergraduate project, publicly presented as a research poster on April 9, 2026.

Future directions

Potential follow-up includes checking the ENC estimator and sequence filters, evaluating alternative explanations, and extending comparisons to additional genomes. These are future directions rather than completed analyses.

Posters & research outputs