This is a small demo codon optimiser that I built out of curiosity. It's also a good opportunity to brush up on the ol' React, more on that here. Expect rough edges as I made this over the weekend. If you try to break this page, you probably will. You can find the code for the optimizer here.

Little disclaimer: This tool is by no means perfect and has some limitations. E.g., it doesn't really validate your input sequence in the sense that it doesn't check if your input "makes sense" (is this even a "valid" protein, does it start and end correctly etc.). It also doesn't check for physical limitations: does this bacterium even have enough tRNAs to build the protein (it should run out at some point, right?). For now, it just assumes that the frequency distribution of tRNAs remains fixed, which is probably incorrect.

Here's my understanding of what a codon optimiser is, based on a recent chat with my dear friend Mr. Claude:

A codon is a group of 3 bases in RNA (and DNA too I guess?). Each of them code for a specific amino acid, which are the building blocks for proteins.

There are 64 codons, but only 20 amino acids, which means multiple codons can code for the same amino acid, e.g. these are all "synonyms" for Leucine: CUU, CUC, CUA, CUG, UUA, UUG.

The cell uses tRNAs (small molecules) to physically carry amino acids to the protein-building machinery (an enzyme I think). Each tRNA recognises a specific codon. BUT: organisms don't have equal amounts of each tRNA type. If your gene uses "CUA" a lot, the Ribosome has to wait for rare tRNAs to become available. Ain't nobody got time for that. You gotta make due with what you have.

For a codon optimiser, you can have as input either a suboptimal DNA string that needs to be optimised. But wth is a suboptimal DNA? Well, let's say you - dear reader - are an E. Coli bacterium and I'm a mad scientist. I will give you a bit of human DNA and you have to make me some insulin. Human codon preferences are different from your codon preferences. You prefer to have CUG codons for Leucine, while the human gene prefers CUA. Now, you're in a pickle.

You are given a foreign bit of DNA that tells you to use CUA, but you can't. You just ran out of CUA tRNAs. You can now either wait (the protein production stalls, perhaps falls off and you get a useless protein) or you make an error (equally useless protein). The solution? Before giving you the DNA, I (using the power of black voodoo magic as far as my understanding goes) rewrite it. I swap out all the CUA codons for CUG codons because I know that those are your favourites. Same amino acid, different spelling. Now when you read it, you breeze through — plenty of CUG tRNAs available. And THAT is the optimisation part: I optimise the given DNA strand to fit YOUR tRNA frequency distribution.