BioDeviceHub

Codon Usage Calculator

Count the codons in a coding sequence and report each codon's frequency per thousand, its relative synonymous codon usage (RSCU) and the GC content at third positions.

Formula

f1000=ncodonntotal×1000f_{1000} = \dfrac{n_{\text{codon}}}{n_{\text{total}}}\times 1000
RSCU=ncodonnˉsynonymous codons\mathrm{RSCU} = \dfrac{n_{\text{codon}}}{\bar{n}_{\text{synonymous codons}}}
GC3=ncodons ending in G or Cntotal×100\mathrm{GC3} = \dfrac{n_{\text{codons ending in G or C}}}{n_{\text{total}}}\times 100
RSCU\mathrm{RSCU}
relative synonymous codon usage: 1 means no preference, above 1 a preferred codon, below 1 an avoided one
GC3\mathrm{GC3}
GC content at the third, usually synonymous, position of the codons

How it works

Most amino acids are encoded by more than one codon, and organisms and genes use the synonymous codons unevenly. Counting how often each codon occurs, and comparing it with the average for its amino acid, shows the bias. The bias matters when a gene is moved between species or synthesised for expression.

GC3 is a compact measure of that bias. It varies widely between genomes, from under 30% to over 80%, and tends to track the genome's overall GC content more closely than the other codon positions do.

Worked example

An illustrative 72-base coding sequence of 24 codons (23 amino acids and the stop codon).

  1. Count each of the 24 codons, for example GAA three times.
  2. Five of the 24 codons end in G or C, so GC3 = 5 / 24 = 20.8%.

24 codons counted, GC3 of 20.8%. With so few codons the table describes this gene, not the organism.

These are the values the calculator opens with, so you can check its output against this example.

Assumptions

  • The sequence is a coding sequence read in frame from its first base, with no introns.
  • The stop codon, if present, is counted as a codon.
  • Codon usage estimated from one short gene is noisy. Reliable species-level tables use many thousands of codons.

Common mistakes

  • Pasting genomic sequence with introns, or starting at the wrong base, which shifts every codon.
  • Comparing RSCU values from genes of only a few dozen codons.
  • Optimising a gene for one host by codon frequency alone. Codon context, mRNA structure and tRNA supply also affect expression.