Filed under: colonizing space, deep time, futurism, Hot Earth Dreams, livable future, Real Science Content, science fiction, Speculation, sustainability | Tags: Interstellar Travel, science fiction, Speculation, sustainability
I came across this little bit when listening to NPR’s On The Media. The episode is entitled “Digital Dark Age” which of course pricked my ears up immediately, as the digital dark age is something I dealt with in Hot Earth Dreams. The whole hour is worth listening to, but the weird idea I wanted to focus on is the idea of using artificially generated DNA for long-term data storage, an idea put forward by Dr. Nate Goldman in this segment.
Superficially, this is a great idea. Dr. Goldman is working on this idea as a way to store the huge amount of genomics data he has to curate at the European Bioinformatics Institute. DNA is pretty stable and information dense, so if it’s possible to cheaply generate long DNA sequences and to cheaply read them, it’s a good form of ROM (Read Only Memory). Dr. Goldman develops this into an idea of caching the great works of civilization in some sort of time capsule that starts by explaining what DNA is and how the code works, then progresses to simple decoding examples, and finally to the whole earth encyclopedia, or whatever is supposed to be in the data cache. DNA is certainly more durable than known electronic digital media and is smaller than durable analog media like baked clay tablets, so superficially it has a lot going for it.
One little problem with this scenario is the idea that it’s easy to generate and read DNA. It’s easy now, but I remember how hard it was even 20 years ago when I was in grad school. This is a new technology. Indeed, Dr. Goldman doesn’t think this technology will be financially viable for another decade or two, although it’s borderline technologically viable now.
Still, DNA ROM works better if we’re talking about a hypothetical sustainable civilization, as opposed to leaving some sort of time capsule for the next civilization 5,000 years from now or whenever. DNA is not the kind of storage medium that will allow people to jump-start civilization from a hidden cache. It’s just too tricky to read and write, even though DNA has demonstrably lasted tens of thousands of years in fossil bones under ideal conditions.
It’s even more suitable when we’re talking about interstellar colonization, where information needs to be stable for thousands of years. Not only can the genomes of potentially useful organisms be stored as DNA, all the other information the starship needs to curate can be stored as DNA as well.
The other little problem with using DNA to store data is that having such technology widely available means that high-level synthetic biology will be available to anybody who wants it. After all, if the equivalent of a laptop can generate as much DNA as your average genome, how many more bits of equipment are needed to twirl that DNA into chromosomes, insert it in a cell, and make a new eukaryotic life form? Letting this kind of technology be available to the public is something that is currently forbidden, at least in current American society. What kind of societal changes would required for people to believe that such technology is safe?
Still, it’s another possible technology for a hypothetical sustainable and starfaring civilization. Perhaps in the future, we’ll have computers that are as much biotech as chips, where spam is something you feed your machine to support its self-repair function, rather than something you delete from your inbox.
Or maybe we should try to baked clay tablet thing…
3 Comments so far
Leave a comment
Putting the life back in science fiction 
If the average base pair has a molecular weight of 650 daltons, and you store 2 bits per base pair, then in 1000 grams of DNA you could store…
( ( ( (1000 / 650) * (6.02 * 10^23) ) * 2) / 8 = 231,538,461,538,461,538,460,265 bytes.
Or in more convenient units, ~196 zetabytes, where a zetabyte is a billion terabytes. For pedants: I am using the powers-of-2 notation common to digital information storage; 196 zebibytes if you want to be explicit.
196 zebibytes per kilogram is roughly 5 billion times as information-dense as current state of the art flash memory, which reaches only tens of terabytes per kilogram; wow!
If you try to keep that kilogram of concentrated DNA in a single container, it forms tangled structures, slowly hydrolyzes and cleaves, forms dimers from light exposure, and certainly can’t be read as a single error-free strand by any known or reasonably conceivable sequencer machine. Now of course error correcting codes come into play here, just as for digital storage. You can increase the robustness a lot at the expense of a modest information overhead with e.g. Reed-Solomon codes. What’s the underlying storage degradation rate that we’re working against? It’s not quantified in this interview, and in the case of bare DNA-inna-flask, like students see in an undergraduate lab, I believe the base rate of information loss is more rapid than with archival digital media.
Goldman talks about storing the DNA in a stabilized form that’s more like dust, but there’s also no quantitative information given about the information density or endurance of the stabilized version. Nor about how much supporting machinery you would need to write new records and read old ones.
My cynical take: this pitch sounds rather like a nuclear cheerleader promoting nuclear power as “handheld” because you can lift a lifetime’s worth of purified fuel in one hand. While ignoring the armies of labor and mountains of ore and infrastructure required to actually make electricity starting from the uranium mine. It’s staggeringly compact as long as you ignore most of the pieces you need for a practical system.
Of course with a factor-of-5-billion raw density advantage there’s a lot of potential for making DNA storage practical and still coming out ahead of conventional digital technologies. But usually people promoting a concept share at least approximate numbers when the numbers are good, stick to vague gee-whiz stuff when the numbers are unimpressive or the project is too immature for quantitative description. This interview read more like vague gee-whiz to me.
Comment by Matt January 9, 2016 @ 12:08 amThanks for putting some numbers to it. I mostly thought it was amusing, to think of a dust-like storage medium and also conceive of it as a way to archive the information of civilization. I guess Eppendorf-size tubes made of, oh, black quartz or something similar, and well sealed with nitrogen gas around the DNA might work for an archive. Still, you’re right, this is a mathematician having fun handwaving, not a serious engineering proposal.
Still, for anyone else reading this if you’re thinking about a hard SF story and want to talk about legitimately storage media, DNA is a way to go. You just have to remember that if your laptop can read and write to DNA, your genome can be edited with technology that looks very similar, so lifehacking will be a real issue for your characters.
Comment by Heteromeles January 9, 2016 @ 12:14 amGreat comment on the quantitative Matt and it stirred some of that dust in my brain left over from Grad school. I did remember that a dalton was an atomic mass unit and was perhaps the weight of a proton or neutron or some such. I just had to look it up and see that it is 10 to the -27 kg.
Comment by cal48koho January 9, 2016 @ 2:07 pmOn a more serious note we need to as a civilization owe a duty to our descendants to have a reliable replicative simple method of information preservation. Clay tablets are hard to beat but paper has worked pretty well and not the sort of acid washed paper we use now which falls to bits after a few decades. We need to be able to preserve critical information and not a mountain of data in a one/zero format. As a practical matter we probably should ban short life paper that goes into anything that is archivable and then of course we need a place to put it, some nice dry cave for example or maybe a salt dome in Louisiana. And then we need to post a guard outside of course.
What was really interesting to me was the astounding information storage capability of DNA which is just the ticket when you are building something complicated like living organisms for example.