Scientists Discover a New Way To Use DNA As a Storage Device

Mark Wilson shares a report from BetaNews: Researchers from the Waterford Institute of Technology (WIT) in Ireland have developed a way to use bacteria to archive up to up to one zettabyte in one gram of DNA. The technique uses double-strained DNA molecules called plasmids to encode data which is stored in the Novablue strain of the E Coli bacteria. The Novablue bacteria has a fixed location, making it viable for storage, and the data can be transferred by releasing a mobile HB101 strain of E Coli which uses a process called conjugation to extract the data. The antibiotics tetracycline and streptomycin are used to control this process. The method is currently not only expensive, but also slow. Data retrieval takes up to three days at the moment, but researchers believe it should be possible to dramatically speed up this process. Equipment already exists that can be used to write to DNA in seconds. Stability and security are also an issue right now, but it is very early days for the technique, and these current downsides are not viewed as being significant enough to write it off. Potential uses for this method of data storage that have been suggested include the recording of medical records in human DNA, and increasing the traceability of the food chain.

T-Virus

[darkain]

Do you want the Resident Evil T-Virus? Because this is how you get the Resident Evil T-Virus.

Re:

[michelcolman]

I wonder if some day we'll finally figure out that our "unused junk DNA" portions actually contain an entire copy of something resembling the hitchhiker's guide to the galaxy.

Re:

[LifesABeach]

I just thinking that finally someone has addressed my storage needs for all my porn. Now if they could just decrease the time to read or write the data.

Re:

[tonique]

If transferring one zettabyte takes three days, you get something like 3 gigabytes per second, which might be enough for your needs.

Re:

[tlhIngan]

If transferring one zettabyte takes three days, you get something like 3 gigabytes per second, which might be enough for your needs.

Meh. That's only about 50-100% faster than what you get with current SSDs (the top end SSDs can do 2200MB/sec reads and 1500+MB/sec writes)

Re:

[Hal_Porter]

An ejaculation releases 250 million sperm. Each sperm contains 3234 Mbp [quora.com]. Each basepair contains 2 bits. Male orgasms take 5-22 seconds on average [glamour.com].

So we can work out the bit rate. (( 250 million * 3234 million * 2bits)/13.5 seconds) in petabytes per second [google.com] = 14.9722222 petabytes per second.

Aww yeah! Bandwidth, baby!

Storing Data No

[wolfheart111]

Generating electricity Yes :) Whole specimens.... :P

Interesting tech all round

[TheZeitgeist]

I recall Harvard guys doing data encoding with DNA a few years ago. Same cost/benefit in their technique (don't recall exact specifics beyond the four bases = quaternary number system in their encoding scheme): The information density is vast, but I/O slooow.

Their application ideas were interesting though; like tiny cameras with memory integrated into wall paint - it would archive history of the room. All kinds of stuff one would think of needing tape backup circa 1995 these guys were pretty sure could be f

This won't end well

[Latent Heat]

"The antibiotics tetracycline and streptomycin are used to control this process"

What happens when generations of the bacteria develop resistance?

Re:

[Anonymous Coward]

These bacteria's are already resistant. The antibiotics are used to keep a selective pressure. A plasmid is a small circular DNA independent of the chromosomes. These are easily lost for example during cell division. To ensure the presence of the plasmid inside the colony of bacteria, the genes of resistance to the antibiotics are inserted in the plasmid (and the data). Cultivating the Bacteria on a antibiotic rich environment only the one having the gene survive. This process ensures that the colony contai

Re:

[sheramil]

I recall Harvard guys doing data encoding with DNA a few years ago. Same cost/benefit in their technique (don't recall exact specifics beyond the four bases = quaternary number system in their encoding scheme): The information density is vast, but I/O slooow.

Oh yeah.

"Data retrieval takes up to three days at the moment, but researchers believe it should be possible to dramatically speed up this process."

I don't think any amount of cache ram is going to help there.

Slow?

[Mr. Sketch]

1 zettabyte = 1e9 terabytes
3 days = 259200 seconds

1 zettabyte / 3 days = 3,858 terabytes per second

That sounds pretty quick to me, considering that the current Fiber Optic speed record [wikipedia.org] is only 1,050 terrabits per second or about 131 terrabytes per second.

Re:

[LifesABeach]

3 days for a response might not be fast enough. Multiple Read/Write devices could reduce the time. My thinking is that we are observing a Virus, not a Bacteria.

Not a biological thing at that scale...

[RyanFenton]

For those wondering, there's been studies and research on maximum genome size, and beyond a certain point (a couple hundred picograms) the mechanisms don't work for copying/processing/maintaining DNA. The largest animal genome would be around 132 micrograms (marbled lungfish)

http://www.genomesize.com/stat... [genomesize.com]

A picogram is 1/1,000,000,000,000 of a gram, so yeah, getting up to a full gram of the stuff, you can store a LOT of raw data - but it's not something you're going to have survive on its own without a very specialized sort of engineered life/system, and likely secondary specialized critters to defend it against regular old microorganisms.

The other problem would be read/write speed - DNA replication isn't fast with current life processes... scale that up a few billion times, and you're definitely going to have to do something different than how life as is does things.

Ryan Fenton

Sue E.Coli for Copyright Infingement?

[aberglas]

If the E.Coli breeds along with its data, who do you sue?

Much more important question than those of mere technical possibility.

Re:

[LifesABeach]

Kind of sounds like an Engineering Problem to solve.

Re:

[LifesABeach]

The journey of a thousand miles begins with a single step. And stop interrupting the grown ups, they are currently working on this problem.

Re:

[LifesABeach]

You interrupted again.

Antibiotics abuse much?

[Bruha]

Seems an incredibly stupid way to use antibiotics.

Obvious speedup, from management

[Snotnose]

If it takes 3 days for a bacteria to produce a datum, then surely 3 bacteria can produce the datum in a day. Add a million or so and you've got yourself a speedy little data retrieval device. It's nothing but an engineering problem.

Future workplace excuses

[jrumney]

Sorry boss, the data storage evolved so I couldn't complete those TPS reports you asked for. And I'm taking the rest of day off because the bacteria in my computer gave me a bad case of the Dreaded Lurgi.

Data Center

[hcs_$reboot]

Have 37.2 TB [smithsonianmag.com] available.

They'll have to be careful with the coding scheme.

[Ungrounded Lightning]

They'll have to be careful with the coding scheme.

Otherwise it would be possible to write an actual virus just by storing appropriate data into a file.

Note that both DNA and RNA can have enzymatic activity - including functions that would cut them out of the backbone and form them into a viral genome, along with the molecular machinery to package and deliver it.

Imagine a Beowulf cluster of these

[romit_icarus]

*ducks*

Error correction

[tigersha]

Can you RAID these things? What sort of error correction exists here? ECC DNA?

Facepalms

[DivineKnight]

Yes, we generally know this, tell us something new. Engineering principles haven't changed -> you can create logical circuits out of plumbing, electricity, optics, what-have-you...the same applies for data storage.

Viral Videos?

[afc_wimbledon]

The obvious first application perhaps??

Good news for mad scientists!

[BeeArt]

In the old days, they had to hide their data in their cat's collar, now they can store it in the cat itself!

Wow

[Travelsonic]

1 zettabyte = 1 trillion gigabytes. That's a lot of data in a small space. If they can make this thing practical, it would really help out in technology sectors that require storage of insanely large quantities of data (like, say, YouTube, to name on example)

Re:

[fafalone]

By the time this technology is commercially available, large companies will still need rooms full of equipment because they now require yottabytes for the 64K SuperHyperHDR 4D Holovideo everyone's iPhone 128 is recording.

Gibson

[cas2000]

Johnny Mnemonic would be a much shittier story if E Coli were involved.

HIT ME!

[Joe_Dragon]

HIT ME!

Optimization

[manu0601]

How do they plan to prevent E. Coli to optimize out this useless DNA?