Rare codon analysis
Flag the codons of a coding sequence that the expression host rarely uses.
Flag the codons of a coding sequence that the expression host rarely uses.
The sequence is read in frame from its first base, three bases at a time. Each codon is looked up in the codon usage table of the chosen host, and codons whose frequency is below the threshold are marked in the sequence and listed in a table under Details. Stop codons are not scored. The summary gives the share of rare codons and every run of two or more rare codons in a row, which are the places most likely to slow translation.
The example is a 25-codon reading frame. For E. coli with the default threshold of 10 per thousand, 24 codons are scored (the final TAA is a stop) and 9 of them are rare: AGA (2.0) at codon 3, AGG (1.1) at 4, CTA (3.8) at 5, CCC (5.5) at 6, GGA (7.9) at 7, TCT (8.4) at 8, ATA (4.2) at 12, CGA (3.5) at 13 and TGC (6.4) at 24. That is 100 × 9 / 24 = 37.5 %. Codons 3–8 form a run of six rare codons and codons 12–13 a run of two. TGC shows the limit of a per-thousand cut-off: it is the preferred cysteine codon (56 % of synonyms) and is only flagged because cysteine is scarce. For yeast the same sequence has 7 rare codons (29.2 %), for human 5 (20.8 %).
The frequencies are those of the Codon Usage Database at the Kazusa DNA Research Institute (Nakamura, Gojobori and Ikemura, Nucleic Acids Res 2000), compiled from GenBank coding sequences:
| Host | Kazusa entry | Coding sequences | Codons |
|---|---|---|---|
| E. coli K-12 | Escherichia coli W3110 | 4,332 | 1,372,057 |
| Yeast | Saccharomyces cerevisiae | 14,411 | 6,534,504 |
| Human | Homo sapiens | 93,487 | 40,662,582 |
W3110 is a K-12 strain, and its genome-wide table is used rather than the older 14-gene K-12 entry. The standard genetic code is assumed; see the codon table. To check the protein that the frame encodes, use theDNA to protein translator.
A codon that the host organism uses seldom in its own genes. Rare codons are usually read by low-abundance tRNAs, so a gene that contains many of them, especially in a row, can be translated slowly, stall the ribosome, or give truncated and misincorporated protein when it is overexpressed.
The classical set is AGG, AGA and CGA (Arg), CTA (Leu), ATA (Ile), and to a lesser degree CGG (Arg), CCC (Pro) and GGA (Gly). In the E. coli K-12 table used here, AGG occurs 1.1 times per thousand codons and AGA 2.0 times. With the threshold lowered to 6 per thousand the tool flags almost exactly this set.
How many times the codon occurs per 1,000 codons in the coding sequences of the host. All 64 values add up to 1,000. It depends on how common the amino acid is as well as on codon preference, so the table also shows each codon as a fraction of its synonymous codons.
Replace them with frequent synonymous codons by site-directed mutagenesis or gene synthesis, or use a host strain that carries extra copies of the rare tRNA genes, for example Rosetta or BL21-CodonPlus for E. coli. Runs of two or more rare codons and rare codons near the start of the gene matter most.
It is a common rule of thumb for flagging codons worth a look. It is deliberately inclusive: both cysteine codons fall under it in E. coli simply because cysteine is a rare amino acid. Check the fraction-of-synonyms column, and lower the threshold to 5 or 6 to see only the codons most likely to limit expression.