Codon Adaptation Index (CAI) Calculator
Score a coding sequence's codon usage against an expression host, before optimising.
🌐 Nothing you paste is logged or stored — every tool is also callable via REST & MCP, and in bulk from the batch tools
Before you codon-optimise, find out how well your sequence already matches the host. Paste a coding sequence and pick an expression organism; the tool computes the Codon Adaptation Index (CAI), the %GC at third positions, and a per-codon adaptiveness chart that highlights rare codons which can stall translation. If the score is low, hand off to the Codon Optimizer in one click.
0 bp
Paste a coding sequence to analyse.
The Codon Adaptation Index (CAI) is the geometric mean of each codon’s relative adaptiveness (w = its frequency ÷ the most-frequent synonymous codon) in the chosen host, excluding Met/Trp and stops. GC3 is the GC content at third codon positions. Run this before deciding whether to codon-optimise.
Working on a whole construct rather than one CDS? Open SeqStudio — the full editor, with live feature annotation, plasmid maps, restriction and primer panels, undo/redo, multi-document tabs and GenBank / SnapGene import and export.
How to use the Codon Adaptation Index tool
- 1Paste a coding sequence (starting at ATG) and choose the expression host.
- 2Read the CAI, GC3 and the per-codon adaptiveness chart (red bars are rare codons).
- 3If needed, click through to the Codon Optimizer to improve it.
Frequently asked questions
What is a good CAI value?
CAI ranges from 0 to 1; higher means codons closer to the host's most-frequent choices. Highly expressed genes often sit around 0.7–1.0, and values below ~0.5 suggest codon optimisation may help — though CAI is only one factor in expression.
How is CAI calculated here?
As the geometric mean of each codon's relative adaptiveness (its frequency divided by the most-frequent synonymous codon) in the chosen host, excluding single-codon amino acids (Met, Trp), stop codons, and any codon containing an ambiguity code, since a codon that could be several different codons has no one relative adaptiveness to contribute. Never-used codons are floored to avoid a zero product.
Why did my sequence come back with no CAI?
Because nothing in the reading frame carried usable information: every codon was Met, Trp, a stop, or ambiguous, all of which are excluded from the geometric mean. Rather than report a meaningless 1.0 from an empty product, the result says so explicitly and reports how many codons were skipped as ambiguous. Check that the frame start is right and that the sequence isn't mostly Ns.
Which organisms are supported?
E. coli, human, yeast, CHO, Pichia, insect (Sf9), Arabidopsis and zebrafish codon-usage tables, the same ones used by the Codon Optimizer.