ToolsAbout

Split codons

Separate a coding sequence into codon positions 1, 2 and 3, or space it as triplets.

Result
What does splitting a sequence into codons do?
A coding sequence is read three bases at a time. This tool cuts it into those triplets and shows them in the two ways people need. Codon positions writes three sequences: all the first bases of the codons, all the second bases, and all the third bases. Spaced tripletswrites the same sequence with a space between codons so the frame can be read by eye and pasted into a figure or an alignment note.
Why separate the codon positions?
  • Codon bias. The third position is where synonymous choice lives, so GC3 is the clearest single measure of codon bias and of how a gene was designed or adapted.
  • Phylogenetics. Third positions saturate over long divergences. Analysing them as a separate partition, or leaving them out, is standard practice.
  • Selection. Comparing changes at the second position, which almost always change the amino acid, with changes at the third, which mostly do not, separates selection from mutation pressure.
  • Quality control. A coding sequence in the right frame has an uneven composition across the three positions. Three near-identical values usually mean the sequence is not coding, or not in frame.
GC at each codon position
GC1 = 100 × (G + C at position 1) ÷ number of codons
GC2 and GC3 the same for positions 2 and 3
GC of the coding sequence = (GC1 + GC2 + GC3) ÷ 3
Worked example. "Try an example" loads the 720 nt coding sequence of EGFP, which is240 codons with nothing left over. The split gives GC1 56.7 %,GC2 31.3 % and GC3 96.7 %, averaging the 61.5 % GC of the whole gene. A GC3 of almost 97 % is the signature of a synthetic, codon-optimised gene: EGFP was rebuilt with the codons preferred by human cells, nearly all of which end in G or C. A natural gene of the same GC content would show a GC3 in the sixties. Set the frame offset to 2 and the values become 31.4 %, 97.1 % and 56.9 %, the same three numbers shifted round, which is what an out-of-frame split looks like.
Reading frame offset
The tool assumes the first base you give it is the first base of a codon. If the sequence begins one or two bases into a codon, set the offset to 2 or 3; the leading bases are dropped and everything is read from there. Bases at the far end that do not fill a codon are left out and counted in the warning under the result. To find the frame in the first place, use theORF finder orsix-frame translation.
Ambiguity codes are kept in place so the frame cannot shift, which is the one thing that matters here. To see the protein, useDNA to protein translation; for codon-by-codon frequencies in a host, rare codon analysis; and for the genetic code itself, the codon table.
Frequently asked questions
What are the three codon positions?
Every codon is three bases: the first, second and third position. Splitting a coding sequence by position gives three sequences, one holding every first base, one every second and one every third, each a third of the length of the original.
What is GC3 and why does it matter?
GC3 is the GC content of the third codon positions. Because most third-position changes do not change the amino acid, GC3 drifts with mutation and selection pressure and is the main signal of codon bias. It is high in codon-optimised genes and in grass and mammalian GC-rich isochores, and it is used to spot horizontally transferred genes.
Why do phylogenetics pipelines remove third positions?
Third positions change fastest, so over long divergences they saturate: the same site mutates repeatedly and the apparent distance stops growing. Removing them, or analysing them as a separate partition, reduces long-branch artefacts. Splitting by codon position is the first step.
What happens to a sequence that is not a multiple of three?
The leftover one or two bases at the end are not a complete codon, so they are left out and reported under the result. If the leftovers look wrong, the sequence probably starts in another frame; change the frame offset.
Are ambiguity codes and gaps kept in place?
IUPAC ambiguity codes such as N are kept where they are, so the frame does not shift. Gap characters are removed before splitting, so run an alignment through a gap cleaner first if the columns have to stay aligned.