Slashdot Mirror
Using Bacterial DNA For Data Storage
NPV writes "January ACM Communications has an article on the use of DNA in genetically modified bacteria to store information. This is an attempt to achieve the ultimate in archival storage (one of the modified bacteria can tolerate 1000X more radiation than a human being). Now just suppose that the "junk DNA" in the human genome is the documentation package for the machine code. Who wrote that manual?" Here's the article abstract.
.... ....
.... ....
Doctor, my stomach hurts!
(1 year later)
Plague Plague Plague!
What we see depends on mainly what we look for. -- John Lubbock Now search for that bug slave!
It's not like copying some extra genetic material is that expensive for the cell. What's the selective disadvantage in having some superfluous introns (non-coding regions) in your DNA?
We may not immediately be able to make natural organisms "better" in terms of natural fitness, but we can still make plenty of modifications which are beneficial to us. We can do it even without the use direct genetic engineering; we call that "domestication".
Java: the COBOL of the new millenium.
Just to be clear, no non-coding segments have been found in bacteria yet (last I heard).
My first impluse was that this is way off. I'm used to working with plasmids where frequently like 60% of the sequence is junk. They use E. Coli and D. radiodurans in the study mentioned in the article. A brief survey of E. Coli K12 (the parent of most common lab strains) sez that about 5-10% of it is non-coding. The old initial reference claims about 11% is non-coding, but a good chunk of that may be regulatory. The radiodurans genome is about 9% non-coding. The up shot is that there is actually a fair amount of 'junk-DNA' in (at least the Coli) bacterial genomes. Not a lot by human standards but enough to be able to squeeze in a chunk here or there if you're careful.
Another impulse was 'gad... that made it into Nature!?' (the journal, the article cited is a self congratulatory summary of their Nature paper). A lot of it follows a well duh kind of reasoning. 'Well duh' science is often the really good kind, but I wasn't particularily amazed by this. The DNA manipulation methods are beyond standard now, the only really clever thing was proposing the use of radiodurans as the host. Even that was sort of obvious (a blazingly well studied organism that is transformable). The DNA -> text using a 6 bit space? Well if you've ever designed linker regions in proteins I'm sure you were at least thought about spelling out you name or something in amino acids (unless your name is BOB). In part this is because every one learns the amino acids by doing stupid things like spelling out their name. Few people actually do this, mind you, as it usually would have some deleterious effect, but the point is I'm sure they weren't the first ones to try something like this, probably just the first to get funded to do this explicitly. Their big addition was to come up with a 3-letter code that includes all the letters and, ooo, punctuation. Then they spelled out bits of 'It's a small world.' My point is that it's not that far fetched and a bit surprising (to me) that it made it to Nature.
As to the utility of these things for information carriers... Mutation would be a problem in the long term. Sure radiodurans would survive nuclear war (these guys put cockroaches to shame) but they do it using lots of mismatch repair and recombinatorial repair methods. These are not perfect repair systems, they can and frequently do introduce many errors, especially in non-essential DNA space. Tying it to a functional protein isn't a bad idea, but unless the added sequence adds some survival advantage it won't enhance the lifetime of the measage (ie. if uncorrputed data gives an advantage then it is statistically less likely to propagate). Also, as you mentioned, the bacterium might notice long chunks (they're using 100 characters here) of useless DNA and excise it. For that kind of text, it might be better to just etch it into stone or something, at least you have some hope of seeing it intact in 2000 years.
Your right about Nature, to me its more New Scientist article (I recall seeing a paper in Biotechniques about encoding text in DNA some 5-6 years ago I think that was for copyright messages)
Mutation may not be too much of a problem as you could reconstruct the data by sequenceing many different strains of the bug (sort of bacterial TCP protocol if the packet is corrupted sequence a different strain)
What I'd like to know is what sort of data would you send? Encoding the data would be a bit of a fiddle.... but extracting the data would be a expensive, soul destroying project, reqireing late 20th early 21st centuary tech and if target decendants have that sort of tech there must be better ways of sending messages./P
Nobody believes in 'junk DNA'. It's a stupid media buzzword. Ask any geneticist, any at all, whether they consider 'junk DNA' to be a misnomer or not. If the unknown equaled "junk", there would be no scientists. Go figure.
This has to be the 434340930493rd article where the presenter considers himself clever because he sees an insight... that everyone else does, too. Give it up. The abstract is interesting, if lacking in news or useful information, but its presentation is nothing but annoying.
The easiest way to disprove the "junk DNA" is to remove the "junk DNA" and see if the organism still works. Take for example a computer program where "junk code" is removed. If the program still runs then the code might not be important. However, the "junk code" could be comment code not removed by the compiler, error checking code (which will not activate unless the program hits an overflow then all heck breaks loose), or even just graphic data which would allow a program to run (but with a corrupted image display).
The basic truth of "junk DNA" is that unless somebody has a "decompile into a higher level language" device then removed code could case all sorts of things to go GOOEY later on when certain conditions are met. Heck, if we look back at the early days of BBS protocols you'd remember the FOO junk padding code at the end of many ZIP files just to compensate for buggy data transmission protocols. That padding allowed a certain amount of send errors at the end of a file to be tolerable while keeping the important parts of the file intact.
"Face it, a nation that maintains a 72% approval rating on George W. Bush is a nation with a very loose grip on reality.