CDS Protein Editor
Edit the protein encoded by an existing CDS and get back a minimally changed DNA sequence.
Learn more
What it does
The CDS protein editor runs translation backwards on DNA you already have. Paste the coding sequence, edit the protein it encodes, and the tool returns a DNA sequence that differs from your original only where it must. Every codon you did not touch is copied exactly, so the restriction map, sequencing primers and cloning junctions you already validated survive the edit, and only the codons for changed residues are back-translated with your chosen host's codon usage. A difference table then labels each site as synonymous, missense, nonsense, stop loss, deletion or insertion. It is used by molecular biologists who need to alter a protein inside an existing construct without re-synthesising the whole gene.
How it works
The tool translates your CDS from the first base in triplets to obtain the starting protein, then compares it with the protein you pasted. Residues that are unchanged keep their original codon, which is what makes the result minimal; residues that differ are back-translated with the host's preferred codon for the new amino acid. If you mark deletions with a hyphen the positions are taken literally, and no alignment is needed. If the length simply does not match, a global Needleman-Wunsch alignment of the two protein sequences locates the difference and the result is flagged as automatically aligned, so you know to check it. Each changed codon is then classified by what happened to its residue. The tool also reports base identity against the original CDS, the GC shift, and re-translates the new CDS as a self-check, so a sequence containing B, Z, U or O (which have no codon) is flagged rather than silently mis-reported.
Worked example
Take a 54 bp CDS starting ATG that encodes MHKAEFVKGGSKAEFVK*. Paste the edited protein MHKAEFVKGGS-AEYVK*, where the hyphen marks a deleted lysine and the F to Y swap is a substitution. In keep-original-codons mode the tool copies seventeen codons exactly as they were, drops AAA at codon 12, and back-translates only the new tyrosine, so the output is 51 bp at 90.7% base identity and the GC content barely moves, 35.2% to 35.3%. The difference table has exactly two rows, one deletion and one missense, with the two altered bases highlighted inside the new codon. Switching to rewrite-all-codons leaves the protein identical but recodes every residue that was not already using the host's favourite codon, and those extra rows are all labelled synonymous.
When to use it
Use this editor when a construct already exists and you need to change what it encodes: repairing a frameshift-free point mutation, swapping a catalytic residue, trimming or extending a tag, removing a stop codon that truncates expression, or recoding a residue that was mis-called in a sequencing read. The keep-original-codons mode is the safe default because it preserves the DNA you have already verified, but the pitfalls are real. Silent changes can still alter mRNA structure, splicing signals or translation speed, so a sequence that is identical in protein is not necessarily identical in behaviour. When you back-translate changed residues, check that no new restriction site appeared in a region you plan to digest, and confirm the new CDS has no unexpected stop. If your edit is large, codon optimization is usually the better tool: past a certain point, preserving the original codons stops being a virtue and starts being a source of rare-codon clusters.
How it compares with other tools
Most online back-translation tools take a whole protein and rebuild the DNA from scratch, which is what you want for a new synthetic gene but is deeply unhelpful when you are editing a plasmid you already hold, because every codon moves and your cloning junctions disappear. Codon optimization tools sit at the other end: they recode a whole CDS toward host usage while keeping the protein, but they still rewrite the entire sequence. This editor is the narrow case in between, preserving the DNA you did not ask to change and showing you the exact bases that moved. GeneFlow, the browser plasmid editor on this site, has a CDS protein editor built into its plasmid view, where it rewrites a feature in place using a single fixed codon table; this standalone version works on any CDS you paste, lets you choose the host table and the rewrite strategy, and reports the full difference map with synonymous and non-synonymous labels you can export as CSV.
FAQ
- How is this different from codon optimization?
- Codon optimization rewrites every codon of a coding sequence toward the host's preferred set, which is what you want for a newly synthesised gene. This editor does the opposite: it starts from the CDS you already have and changes only what your protein edit requires. Every codon you did not touch is copied exactly, so your restriction map, sequencing primers and cloning junctions survive. Use codon optimization when ordering fresh DNA, and this tool when modifying DNA that already exists.
- Can I delete or add residues?
- Yes. Mark a deleted residue with a hyphen, for example MHKAEFVKGGS-AEYVK*, and the codon for that residue is removed from the CDS. Added residues work too: each new residue is back-translated with the host's preferred codon, because there is no original codon to preserve. If you do not use hyphens and the length no longer matches, the tool runs a global protein alignment to locate the difference and flags the result as automatically aligned.
- What do synonymous, missense, nonsense and stop loss mean?
- They describe what happened to the residue at each changed codon. Synonymous or silent means the DNA changed but the amino acid did not, which is what happens in rewrite-all-codons mode. Missense means a different amino acid is now encoded, a substitution. Nonsense means a stop codon appeared, truncating the protein. Stop loss means a stop codon became an amino acid, extending the reading frame. Deletion and insertion mark whole codons that were removed or added.
- Why must the CDS be a multiple of three?
- The tool reads the CDS in triplets from the first base, because that is how the ribosome reads it and how the protein sequence is derived. If the length is not a multiple of three, the final incomplete codon would have to be dropped silently, and everything downstream would shift, so the diff table would misreport where the changes are. Trimming the sequence yourself first keeps the mapping between codons and positions exact.