Every molecular biologist worth her or his guanine knows that the genetic code has built-in redundancy. (Even one or two physicists know this.) That is, there are more nucleotide choices (64 choices when grouped into the 3-nucleotide code words used) than are needed to represent the 20 protein-building amino acids plus Start and Stop instructions used by all cells.

The redundancy helps protect against some mutations (wrong nucleotide in a given slot), but as far as is known, it is not used for computer-style error correction.

If there truly is this freedom of choice without harming the organism, then I see a large business opportunity in the 'personalisation' of DNA messages. As with tattoos (also of questionable taste), the DNA of a person could be tweaked to write little love messages or anything else in some universal DNA ASCII code using the otherwise meaningless nucleotides in the DNA sequence of given genes. These could be enjoyed by friends or lovers, could provide some humourous relief in crime labs, and even for future paleontologists who might be trying to clone a person's DNA from some long-frozen toe bone.
-- sqeaketh the wheel, Jul 25 2012

Genetic code has built-in redundancy http://en.wikipedia.org/wiki/Genetic_code
[sqeaketh the wheel, Jul 25 2012]

DNA windchimes http://reedsong.com/about.html
[xandram, Jul 25 2012]

http://en.wikipedia..._Next_Generation%29 [FlyingToaster, Jul 25 2012]

Okay, I can't pinpoint the *exact* Star Trek episode or movie, but ...

[edit: yes I can <link> ]
-- FlyingToaster, Jul 25 2012


I'm bunning it...[+] for possibilities!!
-- xandram, Jul 25 2012


But how would you know that by changing your DNA you weren't going to cause some strange genetic defect in your future children?
-- hippo, Jul 25 2012


Accepting the unwritten assumption that one is encoding messages in coding sequences, apparently there's little codon bias in humans. So this is fine in that regard. There may be a slight issue with regulation due to altered sense-antisense DNA interactions and similar, but personally I don't think that would be particularly significant.
I reckon the encoding potential of this method is log((64-23)/20)/log2 ~=1 bit per codon, or 0.35bits per bp, on average.

However, why do that at all? Why not just put the message somewhere out of the way in a non-coding region? Not only are you less likely to mess something up, but you can then efficiently encode your sequence using the full information potential of the medium (2 bits per bp).
Easier to encode and integrate the message, and easier to identify the presence of a message and decode it.
-- Loris, Jul 25 2012


//there's little codon bias in humans//

There's little bias in the usage of codons (or at least less bias than in many species), but that's not to say that the choice of codon is completely arbitrary or context-free.
-- MaxwellBuchanan, Jul 25 2012


//Why not just put the message somewhere out of the way in a non-coding region?//
I hate to invoke ignorance to make my point, and undercut the original idea as well, but I have heard that emerging knowledge says the non-coding regions are important for gene regulation or some such. Perhaps it is safer to muck with the coding regions, which we think we understand pretty well rather than the non-coding regions which we don't understand much.
-- sqeaketh the wheel, Jul 25 2012


//There's little bias in the usage of codons (or at least less bias than in many species), but that's not to say that the choice of codon is completely arbitrary or context-free.//

Admittedly I didn't mention all known gotchas (avoiding enzymes containing seleno-cysteine is another thought which just occurred). However I think my comment did make it pretty clear that there were other potential gotchas with the proposed method.

I now think that my estimate of the average coding potential is probably on the low side, since aa with fewer codons are generally the rarer ones.

Anyway, encoding messages CTGATCAAGGAGACCCACATCAGC is far more fun in any case.
-- Loris, Jul 25 2012


//the *exact* Star Trek episode //
Enjoyed reading that plot summary. Thanks!
-- sqeaketh the wheel, Jul 25 2012


//I hate to invoke ignorance to make my point, and undercut the original idea as well, but I have heard that emerging knowledge says the non-coding regions are important for gene regulation or some such. Perhaps it is safer to muck with the coding regions, which we think we understand pretty well rather than the non-coding regions which we don't understand much.//

Well, maybe.

But then again, we don't know everything about coding regions. But we do know that they're important, and they're the bits any stuff we don't know about will be targetting.

And bear in mind that theres a lot more of the junk. Quite a bit of that junk is stuff we're pretty sure isn't of any benefit at all. I'd be much happier splicing in a message to outright replace a transposon[1] inserted somewhere apparently 'out of the way' than I would be trying to hack in a message into pretty much any known host gene.

[1] Transposons are small sequences which encode their own movement around the genome. They're essentially little genetic parasites. Human genomes have millions of copies of some of them.
-- Loris, Jul 25 2012


//encoding messages CTGATCAAGGAGACCCACATCAGC is far more fun//

Fair enough. But TCCTGAATGTGA letters are missing.
-- MaxwellBuchanan, Jul 25 2012



random, halfbakery