Is it believable to store information in live DNA for billion years?

Question

Imagine a civilization that approaches Kardashev II. They are deep into spacefaring, they live on a few less than one-hundred colonized worlds. They've not found evidence of other sapient life, so they believe they are alone in the galaxy.

The unfortunate news is that this civilization is dying. No amount of effort can save them, so they decided to leave a message to those who will come after them. Their scientists concluded that if they will scatter messages that will last one billion years there is a good chance somebody will find them.

Nothing natural lasts that long, so they decided to create self-replicating storage units that would feed on simple elements found on rocky planets. The best place to leave these storage units was near volcanic vents on the floor of water oceans. The storage units were small and robust, designed to survive as long as possible.

The storage unit itself was very clever. Every bit of information was used not only as a part of the message but also as an instruction for how to build some part of the unit. The idea was to prevent the storage units from replicating if they became altered (which would also alter the message) and so the altered units would be destroyed by the flow of time.

Despite the intention, after some time, storage units with changed instructions were able to replicate and survive. But, by this time, the civilization that seeded them was already long gone. So long, that even the remains of their Dyson swarms were indistinguishable from asteroids. Eventually, some of these altered storage units changed so much that they obtained consciousness, and even were able to build their own civilization. They called themselves Humans.

---

The question

How believable it is that somebody will find Maxwell equations and some more during DNA sequencing of some ancient organism living on the bottom of the ocean?

---

EDIT: changed the wording for clarity

Answer 1

I wouldn't normally appear on this site, but that is absolutely believable and theoretically possible in "futurology". The concept simply extends the boundaries of current molecular evolution into protein engineering.

Check my other sites, essentially molecular evolution is a core part of what I do.

The storage units need two components (probably three) to be viable for the design, 'message' to be preserved over vast amounts of time. These are:

1. The thing encoding the replication of the DNA is engineered to minimise mutation and that is absolutely feasible by constructing a 'hyper strong proof reading activity' protein(s) (italics is the precise technical word) and probably large numbers of them. So when the 'message' mutates a set of "mutation repair" proteins kick into action" mopping up the error. This happens and would result in replicating with minimal mutation.
2. Even when the 'message' (DNA) changes, i.e. mutates, it is engineered to immediately lose functionality, thereby unable to replicate the message. That 'message' (DNA) including the mutations then dies. The 'message' that does not mutant persists and replicates, thus the mutations are always lost. It's called purifying selection and yes it absolutely happens.
3. Compensatory mutation theory - this too complicated to explain easily.

However, across the vastness of time, no matter how brilliant the molecular design and protein engineering they couldn't identify every permutation (mutation) that would overcome the purifying selection trap, i.e. mutations start persisting. The first thing the mutated message does is dump is the 'proof reading activity', i.e. the mechanism restricting it from mutating quickly. The 'message' then enters mutation 'free-fall', mutations can occur simultaneously , i.e. within the same instance of time within the same message (and that is very 'dangerous' in this context because of the way compensatory mutations work). Then it's all systems go, and upwards through the complexity of life.

So yeah points 1 to 3 we could see why the original scientists theoreticians and wet-lab were overconfident. The design looks bullet proof, but in DNA ultimately nothing is 'bullet proof'.

Cool.

---

I'd better caveat "mutation free-fall". The pure example of mutation free-fall is cancer, there's no question about that, it's no hold bars. Thats not what I had in mind ... The phenomena where there are sudden bursts of mutations are:

• episodic selection
• periodic selection
• selective sweeps

In contrast to "purifying" or "epistatic" selection, these are systems in rapid selection, they adapt and change. True "free-fall", like cancer, the cell simply mutates itself apart.

Some references for the above 3 processes are given in the notes below. These vast research fields however.

The overconfidence is absolutely correct ... in this area for every success there's been multiple failures. It was thought infectious disease was defeated (yeah right), vaccines got rid of viruses and antibiotics got rid of bacteria ... There have been successes (smallpox was eradicated and probably polio too) but with the benefit of hindsight the original optimism definitely wasn't how its worked out. Those little microbes just keep mutating.

They're a complex issue called "neutral theory" if the "message" is only DNA but if the "message" is stored in the protein sequence that will work.

Caveat 2 Storing a message thats under purifying selection would be some feat, certainly at present, but it's not impossible for future generations to develop this particularly via advances in AI. This is the hardest technological feat BTW ... developing point 1 is absolutely doable now, developing point 2 we couldn't do it at present we are simply not at that level at present. What you are doing is forcing function into a "message" and its a level of computation we simply don't have now.

---

Further comments

What I would look at is the preservation of bacterial 16S sequence. This can be used to build a phylogenetic tree of ALL bacteria. If life is 1 billion years old, then thats how long it's been around. It contains extreme sequence conservation (for 1 billion+ year). 16S is pure DNA its a much easier concept to understand. There are lots of protein genes, e.g. DNA polymerase, that have massive amounts of conservation. This however is complicated because it's about DNA->proteins and it's really hard to grasp if biology wasn't you thing. DNA polymerase was what I had in mind in the above statements and re-engineering that to leave a message and make that message central to its functionality so purifying selection will preserve it. Purifying selection has preserved these genes for however old life is on earth (1 billion years?) and maintained strong sequence conservation.

The issue is what I am proposing is beyond current technology, but it's built on current technology. Without understanding where we are now it's hard to project that into the future.

The key I would focus on is the overconfidence in a futuristic brilliant technological solution and how that can unravel into something completely unintended. The emotional stuff, the arrogance of brilliance. Getting into the detailed technology is going to be hard because it's spanning lots of areas with microbial evolution. For example, generating electricity in coal fired power stations, at the time it would have been brilliant (obviously a long time ago) ... now with global warming that seems a very bad idea and we don't know the final outcome.

Answer 2

The mutation rate in human DNA is round about 1E-4 to 1E-6. That means that in each copying of DNA, there is between 1 in 10,000 and 1 in 1 million errors in copying the sequence.

It means that, if there is no particular restorative process, the error rate is thus.

If there is a process to restore the DNA, it can be retained with extremely few errors. For example, haemoglobin seems to be extremely stable. The differences between the molecule in humans and various other species that have similar blood chemistry is very small. The usual explanation is, any mutations in this chemical tend to produce instant death.

But if a portion of DNA is not active in representing some aspect of the organism, there is no reason to expect it to be protected from the error rate. Indeed, the expected error rate will be the towards the larger end. There is no reason for the more careful error correction in DNA that is not evolutionarily important.

Thus we can expect that each generation will produce about 1 error in 10,000 letters.

How many generations since 1 billion years? Certainly millions.

Thus, any information stored in non-expressed DNA would have massive errors. The expectation is that any such non-expressed DNA would be entirely replaced many times over 1 billion years. No useful information would survive.

Answer 3

I find it unlikely. The fact that humans evolved from these storage units is proof that the medium does not have very good long-term fidelity. Clearly, these units are are unable to maintain a pristine copy of the data they encode. After a billion years, I find it unlikely that the original data would even still exist. It is simply unbelievable that the genetic code of the storage unit could be so stable as to survive with zero errors for a billion years, yet unstable enough to evolve into a highly complex sapient organism. Either the original organism can preserve its own DNA but never evolve into anything else, or allow evolution at the cost of corrupting the message.

Consider two different cases, either that the message DNA does something useful, or it doesn't. If it doesn't do anything useful at all, there isn't any evolutionary reason to keep it intact - individuals with a mutated copy of the message will reproduce just as well as ones with an intact copy, so the message will eventually be corrupted. If the DNA does do something useful, there is more pressure to conserve it, but that doesn't preclude mutations. Evolutionary fitness is very importantly a function of the environment - it doesn't make sense that the same organism could represent the peak of evolutionary fitness across different environments over the course of a billion years. As the environment changes over time, so do evolutionary pressures - the designed organism may have been highly evolutionarily fit at one point, but it would not remain that way unless the environment is totally static.

But even if the message did survive, it's unclear how it could be deciphered - all other organisms on the planet descended from the original storage unit and will have a traceable phylogeny of similar genomes. Simply put, the genetic code just looks like it carries the information of genetic code, not some secondary hidden meaning.

Answer 4

Is it believable? Yes! And no...

We believe that a considerable amount of human DNA is dormant. It's the remains of, for example, old evolutionary changes that are no longer useful or viable, or evolutionary traits that never caught hold and were replaced by something else. We also believe that some of it is "how to build a human" DNA that has no meaning once the human has reached maturity.

That inefficiency in our genetic code is due to nature figuring us out on the fly over millions of years.

But what if an intelligent species that understood the inner workings of DNA (much better than we do today...) were to use that knowledge to create an engineered life form that could carry within the dormant DNA space useful information?

Yeah... I can believe that. As I mentioned in my comment, the basics of this idea have already found their way into Science Fiction in the form of the Ultimate Computer in Douglas Adams' *The Hitchhiker's Guide to the Galaxy" series and Star Trek: The Next Generation's episode "The Chase" where multiple species find chunks of "something useful" within their DNA and, when all those chunks are brought together, it forms a program that results in a hologram that would have been uber cool had it not been a huge disappointment.

But the idea requires ignoring a few things to be believable

We humans are remarkably adept at ignoring all kinds of truth in the pursuit of a good story. Look at politics anywhere in the world for proof. However, for the record, the weakness in this idea is that evolution is constantly happening, and the data is nothing more than "unused space" in terms of the evolutionary record that would (over millions and billions of years) become corrupt — guaranteed. The problem is that evolution is as capable of discarding something it doesn't deem useful as it is incorporating it (if you don't mind my anthropomorphising a concept).

Yes, we can say, "but the original encoding contains instructions to avoid that!" But the problem there is that you can't end up with a human in that case. You had to start with one that changed very little over the eons — a bit like alligators, only worse. You can't allow the entire evolutionary process from primordial goo to pat-my-head-and-rub-my-belly human and still "believably" retain all that information.

But who cares?

I like the idea and believe it hasn't been explored to its full depth. I believe DNA is complex enough and still mysterious enough that everyone other than an actual geneticist would happily read your story and accept the premise.

And if you do a good enough job, even the geneticists will enjoy themselves... all the while wishing they could figure out how to do what you're describing.

Answer 5

Whoever wove that Message into our DNA meant it to last until we figured out how to read it—it's built inextricably into the protein expression mechanism so any organism with a corrupted copy won't be viable. It's obvious in retrospect. If a visitor wants to leave a message, why not make it a self-reproducing, error-correcting message that any sentient race would stumble upon as soon as they undertook to reverse engineer their own design?

Source: https://www.fourmilab.ch/documents/sftriple/gpic.html

Answer 6

The storage unit itself was very clever. Every bit of information was used not only as a part of the message but also as an instruction for how to build some part of the unit. The idea was to prevent the storage units from replicating if they became altered (which would also alter the message) and so the altered units would be destroyed by the flow of time.

This is a very good start. In reality, the key parts of the ribosomes responsible for reading mRNA strands and assembling the corresponding proteins is extremely well conserved for exactly the reason stated: any major change is likely to result in different proteins being constructed or a failure to construct proteins at all, very likely making the cell non-viable.

The ribosome itself is composed of ribosome RNA bundled around a "scaffold" of various ribosome proteins. Some parts are actually pretty robust to change, as long as they don't cause any changes to transcription, but the rest of it doesn't tolerate much change at all. The ribosomes of bacteria are all very similar, and those of eukaryotes (except those in the mitochondria, which are similar to bacterial ribosomes) and archaea are very similar, and distinct in the same ways from those of bacteria ribosomes. Either it is the prokaryotes (bacteria) which have diverged, or the archaea did so before eukaryotes appeared.

However...all this is coded for with just a few thousand base pairs of DNA. The smallest known genome is of a bacteria with only 1.3 million base pairs, and only a fraction of that would be the highly-conserved cellular machinery that could store a message, and only a fraction of that would actually be available for storing the message. Compression would make it more likely for some error to make the message unreadable, what you'd want is instead an error-correcting encoding which will incorporate some redundancy in exchange for error tolerance. In short, the message length would probably be limited to a few kilobytes, if that.

So, Maxwell's equations? Maybe. Maxwell's equations and enough contextual information to make them actually recognizable? That'd be less believable. The location of a time capsule stashed deep inside some geologically inert lump of rock in some very long-term-stable orbit? Maybe.

Answer 7

Yes, highly plausible.

Real DNA mutation rates are too high for this to work; however, that's because mutation rate are themselves under evolutionary selection for the optimal balance of energy costs of the repair mechanisms as well as the long term advantage of the organism - i.e. organisms with very low mutation rates evolve very slowly and will eventually be outcompeted, organisms with very high mutation rates degrade in fitness too quickly for natural selection to weed out detrimental mutations.

A manufactured organism could have much lower mutation rates than any extant organism, by having more repair mechanisms, and stricter proof-reading mechanisms and more redundant storage.

Answer 8

Frame challenge

There are two main problems your dying civilisation needs to consider while sending out a DNA message:

1. Any civilisation capable of doing all this will be well aware of DNA's mutation characteristics and how relying on natural selection to only preserve flawless copies is a fool's errand.

2. Sending a message only makes sense if you assume someone is going to read it. Anyone able to decode your message hidden in DNA will be well aware of Maxwell's equations, as they're kind of necessary to build the technology to sequence DNA. You need to put much more advanced information into it. But what if it is found by someone horrible? The more advanced the knowledge encoded is, the bigger the danger of a random recipient misusing it.

So the whole thing doesn't make sense.

Unless...

Why not use the message to build the recipient?

You know those self-extracting .exe files that were all the rage in the early 2000s? They were self-extracting because they contained not just the compressed data but also the code to decompress it.

So far from it being an accident, the mutation of (some of) the message into a sentient species was the aim. A sentient species capable of figuring out Maxwell's equations and analysing DNA, determining its "age" and with a deep obsession of finding out where they came from.

A species capable of finding a multitude of the most ancient copies of your message (encoding stuff much more complicated than Maxwell's equations), find and "correct" the inevitable errors that would creep in over these time scales and figure out what the "starting" DNA might have been.

Answer 9

There are some challenges with storing information in DNA. DNA is not a very stable molecule over time. The Max Plank Institute had to invent new ways of reading DNA that was only 40,000 years old. The DNA was highly contaminated with DNA from other organisms. The DNA strands started fraying. And chemicals modified the DNA.

In our cells, DNA is stored behind two different walls, the cell wall and then the wall around the nucleus in order to protect it.

To have information stay encoded in DNA for millions of years, the DNA would need not just the information, but also a way to look for and fix errors. Something like how NASA uses three different computers at times in order to have enough error detection and correction. Perhaps the organism might have 3 different nuclei in each cell with error correction.

It would also help to have the information duplicated elsewhere so that a catastrophe in one location won't destroy the information. Then, you will need a way to compare stored information between the various locations and take the one that has the most votes.

Think about it. DNA has a number of ways to be destroyed. Another organism eats the cell. Chemical processes eating away at the molecule. Radiation changes the atom in a molecule which makes the information unreadable. Many other types of natural disasters will destroy the information including landslides, volcanoes, and plate tectonics subducting where this information is stored.

Edit: Storing information on the ocean floor doesn't last. Earth does not have any ocean floor older than 200 million years. Everything older than that has been subducted. Storing on land has to contend with both critters eating it and with all the processes that build dirt and rock on top of it or erode away where it is stored. In short, the DNA has to be able to move.