How the genetic code works.

Three bases make one codon

A nucleotide sequence is read in groups of three bases during translation. Each group is a codon. RNA has four usual bases, A, U, G and C, giving 4 × 4 × 4 = 64 possible three-base combinations. In the standard genetic code, 61 combinations specify amino acids and three are termination signals. The interactive codon chart shows the complete mapping, while the codon wheel arranges the same information from the center outward.

Degeneracy of the genetic code

Degeneracy means that several codons can specify the same amino acid. In the standard code, 61 amino-acid codons map to 20 amino acids. For example, GAA and GAG both specify glutamic acid (Glu, E). These are synonymous codons: changing the third base from A to G leaves the amino-acid assignment unchanged. The NCBI standard-code table supplies these assignments.

The one-base explorer above shows every single-base substitution for a codon. There are three positions and three possible replacements at each position, giving nine alternatives. For GAA, one is synonymous, seven specify a different amino acid and one becomes a stop codon. These counts describe possibilities, not their probabilities in a cell.

Redundancy is not confined to the third base. For example, the first-base change from AGA to CGA preserves arginine. Species also favor different synonymous codons; the codon usage tables compare E. coli, yeast and human genes.

What the substitution labels mean

The examples below use the standard code and show both mRNA and coding DNA, written 5′ to 3′. Coding DNA has T where mRNA has U. A template strand must be converted with its direction specified before this comparison.

These labels follow NHGRI's missense and nonsense definitions and the NLM definition of a synonymous or silent mutation. Here, “synonymous” refers only to the amino-acid assignment. It does not establish that a change is biologically harmless. Stops have no amino-acid assignment, so a change between two stop codons is labeled separately.

The explorer uses NCBI table 1 and internal codon assignments. It does not decide whether a codon starts translation or predict splicing, recoding or biological effects. It also does not model insertions or deletions: replacing one base keeps the codon length at three. Use the DNA-to-protein translator to inspect a longer sequence and its reading frames.

Codon versus anticodon

An mRNA codon supplies the triplet you look up in this chart. An anticodon is the three-base sequence in a transfer RNA (tRNA) that pairs with a codon, running in the opposite direction. NHGRI's anticodon definition describes this relationship.

Read AUG, not UAC or CAU, in an mRNA codon chart to find methionine. Select an exact codon in the codon chart and its lookup shows the Watson–Crick anticodon both ways when it encodes an amino acid. A definite stop has none, because release factors, not tRNAs, read it. The anticodon finder pairs every codon of a sequence and has an anticodon chart of all 64. This table is a base-pairing example, not a prediction of every tRNA a cell uses. The strand-direction worked example explains the same orientation issue for DNA.

Some selected tables permit both an amino acid and termination for the same exact triplet. In NCBI table 27, for example, UGA can mean tryptophan or stop. The lookup qualifies its pairing with If read as tryptophan and keeps the stop outcome visible. It cannot choose the biological outcome from the triplet alone. Ambiguous input such as GCN has several possible codons, so the lookup gives no single anticodon.

Wobble pairing at the third base

A codon's first two bases pair strictly with the anticodon. The third pairs more loosely, which Francis Crick called wobble in 1966. In his rules, the base at the anticodon's 5′ end decides which third bases it accepts:

So an anticodon written 3′-GAU-5′ (5′-UAG-3′) pairs with CUA and, through a G·U wobble pair, with CUG as well; both are leucine codons. One tRNA can therefore read more than one codon for its amino acid, which is why codons for the same amino acid mostly share their first two bases: the blocks you see in the codon chart. OpenStax Biology 2e describes the same effect.

The anticodons in the codon lookup are the exact Watson–Crick partners, the pairing these rules start from. Which tRNAs an organism has, and how their bases are modified, differs between species; the anticodon finder lists the tRNA genes of human and E. coli cells that pair with each codon.

The reading frame changes the groups

Consider a sequence that begins ATGAAACCC. Reading from the first base produces ATG, AAA and CCC. Starting one base later produces TGA and AAC, with two bases left over. Starting two bases later produces GAA and ACC, with one base left over. These are the three forward reading frames. Each groups the same letters differently and can give a different protein result.

This table is a codon-by-codon lookup. An ordinary translating ribosome would terminate at the stop in frame +2; the N shown afterward is a separate lookup, not a claim that translation continues through the stop. You can follow each triplet using the standard DNA codon table.

Double-stranded DNA also has the opposite strand. To describe that strand in the usual 5′ to 3′ direction, calculate the reverse complement, then consider its three reading frames. This gives six possible frames in total. The DNA-to-protein translator can show them together. Its complete ORF inspector separately searches for start-to-stop candidates. Selecting a frame is an assumption about how the sequence is read, not evidence that the frame represents an expressed gene.

Starts and stops depend on more than a lookup

AUG specifies methionine in the standard code and is a familiar initiation codon. Initiation, however, requires the appropriate biological context. An internal AUG is generally a methionine codon, rather than the start of a new protein. In eukaryotes, ribosomes scan from the mRNA's 5′ end and the bases around each AUG decide how well it is recognized, summarized by the Kozak sequence. In bacteria, a Shine-Dalgarno sequence a few bases before the start codon pairs with the ribosome's 16S rRNA. Some organisms also use alternative initiation codons. NCBI's tables distinguish ordinary amino-acid assignments from the codons recorded as possible starts.

The usual standard-code termination signals are UAA, UAG and UGA. A stop terminates ordinary translation rather than adding a standard amino-acid residue. The start and stop codons reference lists them for all 27 NCBI codes and counts which ones real genes use. An asterisk is a common representation in translated output. Special biological recoding can change the interpretation of particular codons, so a simple sequence-only translator should state its assumptions. This site's translator applies table assignments and does not infer specialized recoding from surrounding signals.

The genetic code has variants

The standard code is widely used, but it is not universal. Mitochondria and several other biological systems use alternative assignments. For example, UGA specifies tryptophan in the vertebrate mitochondrial code, whereas it is a stop in the standard code. The correct table depends on the organism and the sequence's genomic compartment. Choosing a table because it gives a longer translation is not a substitute for identifying the sequence's origin.

NCBI's genetic-code reference provides the 27 tables available in this site's selector. Compare these internal assignments from tables 1 and 2:

Other tables contain codons that can specify an amino acid or termination depending on context. The charts mark those assignments explicitly. The translator uses X where the input and selected table cannot uniquely establish an amino acid, including these context-dependent cases.

Ambiguous bases do not always mean ambiguous residues

Sequence data can contain IUPAC ambiguity letters. N means any nucleotide, while R means A or G. A codon containing ambiguity can sometimes still resolve to a single amino acid. GCN gives alanine under the standard code because GCA, GCC, GCG and GCT all encode alanine. By contrast, a set of possible codons with different assignments cannot be reduced to one residue without additional information.

Try these triplets in the codon lookup. A mixed result is reported as X when translating a sequence; the individual possibilities remain visible in the inspector.

The reverse complement tool preserves ambiguity by using the corresponding IUPAC pairing rules. The amino acid chart explains the separate one-letter protein alphabet. Keep those alphabets distinct: the same letter can have different meanings in nucleotide and protein sequences.

A useful checking order

Establish the sequence's origin, strand direction, reading frame and translation table. Check whether the sequence is complete and whether the annotation describes a coding sequence. Then inspect internal stops, incomplete codons and ambiguous positions. These checks make a computed translation interpretable. They do not establish expression, protein function or the success of a laboratory experiment.

See sources and methods for data provenance and the exact behavior of these tools.