Codon Optimization & Back-Translation
Optimize a coding sequence for a host or reverse-translate a protein into host-tuned DNA.
Learn more
What it does
This tool does two related jobs. In protein mode it reverse-translates an amino acid sequence into a DNA coding sequence for a chosen host, picking the most frequent synonymous codon for each residue. In DNA mode it rewrites an existing coding sequence so its codon usage matches the host bias while keeping the protein unchanged. Both modes report GC content, the share of codons that changed, and a full codon-by-codon map. It is used by synthetic biologists, protein expressers and students building clones for expression.
How it works
Reverse translation looks up each amino acid in the host codon table and selects the highest-frequency synonymous codon; for degenerate amino acids this pushes the sequence toward the host's preferred usage. Optimization of an existing CDS recodes each codon toward those high-frequency choices and can also silently mutate away from any restriction sites you list, again without changing the amino acid. The output GC percent is the fraction of G plus C over the whole coding sequence, and the substitution percent is the count of codons that differ from the input divided by total codons. In DNA mode the optional avoid list removes unwanted sites by synonymous substitution.
Worked example
Reverse-translate a 73 residue protein for E. coli. Leucine, encoded by six codons, is placed as CTG (the E. coli favorite, about 51% of leucine uses) instead of TTA (under 2%). Arginine becomes CGT rather than CGG or AGA. The resulting CDS shifts toward roughly 55% GC, versus about 45% if every codon were the rare one, and the substitution percent would be near 100% for a fresh reverse translation. If you instead optimize an existing CDS, substitution percent shows only the codons that were changed, often 20-40% for a human gene moved into E. coli.
When to use it
Use reverse translation when you have a protein and need DNA to order, and optimization when a gene expresses poorly in a new host. The main pitfall is treating codon adaptation index (CAI) as the whole story: a high-CAI sequence can still fold badly, carry hidden restriction sites, or repeat a tRNA so often that it stalls translation. Avoid pushing GC to the extreme, since very GC-rich DNA is hard to clone and amplify. For toxic or repetitive proteins, keep some rare codons to slow translation and improve folding. Always check that the optimized CDS has no new stop codons or unwanted sites.
FAQ
- What is reverse translation?
- Reverse translation turns a protein back into DNA by replacing each amino acid with a codon. Because most amino acids are encoded by several synonymous codons, the tool picks the highest-frequency codon for your chosen host, for example E. coli prefers CTG for leucine (about 51% of leucine uses) over TTA (under 2%). The result is a coding sequence biased toward that host's usage, ready to order as gene synthesis.
- What does codon optimization change?
- Optimization rewrites an existing coding sequence so its codons match the host's preferred set, keeping every amino acid identical. It reports the substitution percent, the share of codons changed, and the final GC percent of the coding sequence. In DNA mode you can also list restriction sites to remove by silent mutation. The protein sequence is never altered, only the DNA spelling of each codon.
- Why not maximize the codon adaptation index?
- A very high CAI forces the most frequent codon everywhere, which can exhaust a single tRNA and stall translation, misfold the protein, or create repetitive DNA that is hard to clone. Real highly expressed genes keep some rare codons to pace elongation. Aim for moderate optimization, leave occasional rare codons in repetitive regions, and check that GC content stays in a clonable range rather than chasing a perfect CAI.
- Which host tables are available?
- The host list is built from the bundled codon tables, starting with E. coli, the workhorse for expression, followed by yeast (Saccharomyces cerevisiae), human and others. Each table carries that organism's measured codon frequencies, so the same protein yields a different optimized DNA for each host. Choose the organism you will actually express in; optimizing for E. coli then cloning into mammalian cells gives the wrong bias.