h a l f b a k e r yGood ideas at the time.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Encode a file using the genetic code, so each nybble is represented by a specific codon. Proceed to construct an oligonucleotide as follows: Take some polystyrene beads and bind phosphoramidites to them. Build up oligonucleotides in the usual way, by pulling off the dimethoxytrityl and the group
on the phosphate bit with trichloroacetic acid in a dimethyl sulphoxide solution, then wash it out, add another nucleotide, acetylate the ones which don't work, wash it out, add iodine and water and so on, adding the appropriate nucleotides for the codes via computer, possibly a lab on a chip. Mark each end of the data with a repetitive sequence of base pairs at each end to make it stand out. Add a green fluorescent protein gene and an identifying string which gives a name to the data. Stick it in a yeast artificial chromosome and insert it in a yeast cell. Let them multiply for a while, then dry them, release some as spores, say through an air conditioning unit in an international airport, getting them to spread throughout the world. In the meantime, post the yeast samples to various trusted people.
At the "other" end, use a UV light to find the yeast cells. Culture them in a sugar solution, then add salt solution and detergent, cold isopropanol and yank out the DNA strands with a glass rod. Separate the relevant DNA bits with restriction enzymes which snip off the DNA at the long repetitive sequences. Use a commercial chain-termination DNA sequencing kit to get the information out and convert it back to binary. Do it twice to eliminate mutations and copying errors.
Do this to publicise sensitive information which certain organisations don't want the general public to know about, or simply to distribute open source information, like the latest release of Linux or, more likely, something one hell of a lot smaller.
Go on then, flame me.
Bakedish
http://io9.com/5699...ia?skyline=true&s=i
Done, apparently. [nineteenthly, Nov 27 2010]
[link]
|
| |
So what are you saying? Encode stuff in yeast DNA, so that everyone's bread contains top-secret information? |
|
| |
Partly, but only if it's sourdough (so we're talking Candida rather than Saccharomyces). It would also enable one to say "in vino veritas" with utter candour. |
|
| |
While you wait for the flames, I have some requests for clarification: |
|
| |
//each nybble is represented by a specific codon// Now, a nybble is four bits holding, collectively, one of 16 possible values, right? And each nucleotide is of one of four possible types, so a sequence of *two* nucleotides holds the equivalent amount of information to one nybble (because 2^4 = 16 = 4^2)? Now, if a codon is a sequence of *three* nucleotides (thank you, wikipedia), that means you've got some redundancy there - not a criticism; I'm just checking that I'm following correctly. |
|
| |
//Build up oligonucleotides in the usual way// [marked-for-tagline] |
|
| |
Now, because you're starting with oligonucleotides and building up, I take it that your not talking about using your *own* DNA as a sort of signature ... actually, that would defeat the whistle-blowing point, wouldn't it. D'oh! Sorry, carry on. |
|
| |
//Mark each end of the data with a repetitive sequence of base pairs at each end to make it stand out.// Small dittography there - you probably don't need both of those 'each end's. Just a suggestion. |
|
| |
//a green fluorescent protein gene// a gene which expresses a protein, such that the protein flouresces green, right? |
|
| |
//an identifying string which gives a name to the data// a sort of DNA XML tag? |
|
| |
I'm imagining that you're adding these things just within the repetitive end-markers? |
|
| |
//insert it in a yeast cell// Ah! Now I see why you're expecting flames. You're going to put what is, from a biological point of view, random genetic data into living yeast cells, thereby adding a whole new level of anxiety to persistent yeast infections. |
|
| |
//use a UV light to find the yeast cells// So, I've just got home from the airport, and I suspect that I may have got some of this stuff on me. So, I take off my travelling clothes and, having put on my most fetching boxer shorts, shine a UV light on the former. What am I looking for? Does the green fluorescence you mentioned earlier have anything to do with this? |
|
| |
I'll use my sugar solution to sweeten my tea while I await your answers. |
|
| |
Hmmm. There's no reason why this couldn't work, in
theory. |
|
| |
In practice:
(a) Yeast is a pain in the arse to work with, and I'm not
sure your simple detergent lysis will work. Easier in any
case to use a harmless E. coli strain (easier to clone into,
easier to grow, easier to recover). |
|
| |
(b) Oligo sysnthesis is tricky over about a hundred bases
so, to encode anything longer than a short message, you'll
need to do some more complex construction (perfectly
feasible, but very tedious). |
|
| |
(c) No need to go through the DNA prep. Just put unique,
known sequence at each end of the message, then PCR it
out. (You don't need to do a fancy DNA prep for this; just
toothpicking a colony into a PCR reaction works fine). |
|
| |
(d) The bugs will, during growth, tend to spit out the
message, though probably at low frequency. |
|
| |
(e) probably simpler to print the message on pieces of
confetti and put these into the air ducts. Or broadcast it
on Radio 4's wavelength from a mobile pirate station. |
|
| |
Hello, [pertinax] and [MB]. My idea is to use a specific nucleotide sequence per nybble. Several codons code for each amino acid, but i would prefer to use a specific codon for each hexadecimal digit. There would be redundancy, but not in this case.
Oligonucleotides - "...muss man schweigen" i think.
Yes, it was dittography but at least appropriate because you could encode the word "each end" as delimiters and separate it with "each end" recognising (pathetic fallacy, sorry) enzymes.
Yeast cell - well, it would be non-coding DNA but i don't know how to ensure that it doesn't encode. Maybe if it occurs between 'capital letter' and 'full stop' codons? And yes, XML indeed. And again yes, that's why it's fluorescent.
[MB]:
a) I thought about E. coli, but in an earlier incarnation of this idea i was considering just using non-coding DNA on the actual genome, and i understand the redundancy of E. coli is smaller than that of a eukaryote. I also thought about putting it on tumour cells in the same way. The very earliest incarnation of the idea was to encode info on sperm and get it to manifest after fertilisation in some way, but use it as a simultaneously very slow and very fast broadband connection. Clearly that would be silly.
(b) Then again, the oligonucleotide thing made me think the likes of a mathematics tutorial video or Linux is not feasibly storable in this way, but what would be storable is a message like the ingredients in the Colonel's secret recipe, text-based evidence for Primark's secret sweatshop or Lambeth council's pretend public libraries.
(c) I was concerned about mutation garbling the message and the delimiters, which is why i thought two genomes should be compared.
(d) Could you run that by me again? I was thinking maybe a barcode generated by gel electrophoresis might work if instead of relying on codons, there was just a series of sequences of different length. Easier to read but maybe harder to arrange.
(e) Ah, the Vrillon scenario! Well, if the info is out there in the environment replicating, cleaning up or hushing up strategies are harder. Then again, memes are self-replicating too. |
|
| |
[19thly] (a) You could use E. coli. It'll gladly accept a large
chunk of non-coding DNA (up to a few hundred kb). You
just need to ensure that your message doesn't include the
signals necessary for transcription in E. coli |
|
| |
(b) in theory you could encode any kind of data that can
be represented digitally, within the size constraints. |
|
| |
(c) Well, then just PCR from a couple of independent
colonies (actually, if it's a really long message you couldn't
just PCR it out; but then if it's longer than about a kilobase
you'll have to sub-clone it in order to sequence it anyway,
or do some other such trick) |
|
| |
(d) I just meant that the bugs would tend to lose the
unnecessary DNA over generations. |
|
| |
(e) Actually, the even more reliable way to get information
into the public domain is to entrust it to a government
official with a laptop. |
|
| |
(a) So: That means you could put bits of the Bible on one bacterium, bits of the Origin of Species on another, turn it into coding DNA and see which one survived longer, thereby providing completely spurious evidence for which one is true.
(b) You could do OpenDOS or CP/M then, but not Linux.
(c) Sort of like Bit Torrent or for that matter TCP/IP that.
(d) Yes, hence the need for several different organisms.
(e) If they did that with mine they'd just get Peter Greenaway films and the Indo-Pacific Journal of Phenomenology. |
|
| |
1)In the real world, they can't just multiply endlessly like they do on a petri dish. They'll have to compete with natural yeast. |
|
| |
2)Wouldn't you have to announce to everyone that the information was there, thus revealing yourself to be the source of the information? |
|
| |
3)Its pretty expensive to read so not many people will read it even if you tell them. |
|
| |
baked in quite a few SF novels. |
|
| |
[Jim], (1)they do stand a good chance of dying, but i don't think their fitness relative to wild yeast is affected unless the need to replicate an extra chromosome influences that. If it does, that might be avoidable by using non-coding DNA on the original chromosomes rather than an artificial one. The issue would presumably then be whether more energy is needed for the cell to synthesise some of the nucleotides than others. If that's not feasible, culture the yeast in otherwise sterile conditions (so Saccharomyces rather than Candida again) and produce a block of the stuff containing the right DNA.
(2) No. Just the idea that there is fluorescent yeast containing secrets would be enough. Publish the information on how to do it in detail publically, possibly with a yeast copy itself. You could also ask the organisation to clean up its act before you released it. Clearly the possibility of blackmail exists here, but you can murder someone with a scalpel as well as save their life, so that isn't an argument against the technique as such. The very fact that the technique existed could act as a deterrant. Maybe even the _rumour_ that it was possible would be enough.
(3) I may be wrong, but even if it's expensive now i think a sort of Moore's Law operates, making it exponentially cheaper to do in a short period of time. In fact, aren't there genetic engineering hobbyists even now?
[FlyingToaster] really? I must be out of touch then. The whistleblowing via the internet and the like is there, and i imagine the text via DNA is, but both together? Tell me more please. |
|
| |
// I don't think their fitness relative to wild yeast is affected etc // |
|
| |
If there are a million yeast in your petri dish and a quintillion yeast in the wild, there is a 1/trillion chance that a yeast found a year later will be descended from the petri dish. The million or so encoded yeast that we can statistically expect will not be scattered everywhere either, they'll be clustered together in a few random spots. |
|
| |
It only really works if we do some genetic engineering to make our yeast fitter than the wild yeast. Which could be very hard even if we understood genetics. |
|
| |
Sorry about editing my post btw. |
|
| |
I agree that the yeast would be totally lost in the wild, but it wouldn't be in the wild. It would be on various people's bodies, luggage, clothes and the like. Sent out in a bottle of wine, it would be clearly noticeable as a fluorescent vintage is rather rare. The human eye is sensitive enough to pick up a very small number of photons. Sitting in a darkened club somewhere, a "luminous" alcoholic beverage picking up ultraviolet light would be rather obvious, for example. At the same time, it could be distributed in various other ways: in the wild (hard to find but still there somewhere), on bilberries or other yeasty fruit, in blocks of fresh or dried yeast, and possibly in a loaf though i think maybe the heat would have destroyed the message. |
|
| |
If it was Candida albicans, it would be commensal and fluorescent thrush would be rather obvious. Actually, maybe the best thing would be to use Malassezia furfur, since the differential diagnosis of pityriasis versicolor would probably involve using a Wood's Light, i.e. UV. |
|
| |
This proposal brings up a question which has no been answered - how well do enigneered organisms persist in the environment. The place to look is fruit fluies, which are engineered like mad, and retain the ability to escape and mingle with the general fruit fly populace. The signs of engineering should be pretty clear. Given a banana trap, how far away from the fly lab can you still find flies bearing signs of genetic modification? Down the hall? Outside in the dumpster? Down the block? At the city dump? |
|
| |
//how well do engineered organisms persist in the environment// |
|
| |
Depends on what you've been doing. If you've only put vital information in the junk DNA, about the same as wild equivalents. If you've made it glow, it will use energy that a wild organism would find a use for, so it will not compete in the long run. |
|
| |
You can give them advantages in the same way you can make disease resistant crops. These may do well, but you cannot guarantee success. Random alterations that do not aid an organism will probably harm it. |
|
| |
One possible solution is finding an organism in a foreign country and hope it's invasive like the cain toad or kudzu. If you did it with lots of species you might get a good one. |
|
| |
Lets create a wikipedia-encoded post apocalyptic butterfly species. And lets also create a species that encodes the commercial jingle "Byyyy mennen", because we don't want to forget that. |
|
| |
// Lets create a wikipedia-encoded post apocalyptic butterfly species // - which is uneditable and therefore fails to omit the error which led to the apocalypse. |
|
| |
CATGAACTGC CGATTATGAC CCGCGCGCCG CCGGAAGATA TTAACGCGGG CGAAAACGAA
ACCATTTGCA GCCTGGCGCC GGCGAACGAT TGCGCGAACA CCGGCGAAAC CCAGGTGACC
TAA |
|
| |
Hmm, this will take a while... |
|
| |