21 comments

[ 3.2 ms ] story [ 59.5 ms ] thread
The title suggests that this is an entirely new achievement. But (as also mentioned in the article), this has been possible for several years, since 2012 [1]. Steady progress is being made at decreasing the cost. But the procedure itself isn’t new.

Furthermore, the title suggests that existing DNA molecules are “tweaked” to encode arbitrary information. While that’s possible [2], the procedures in question don’t tweak existing DNA, they synthesise new DNA molecules, which is vastly more efficient.

[1] https://www.nature.com/articles/nature11875

[2] https://www.nature.com/articles/nature23017

Exactly, this DNA does not get to live inside a cell.

The next evolution is to craft viruses that carry the payload and use a (human) host to carry the information around.

I hope they put some DEP in there. I guess that your liver "running" some vacation videos won't be good for your health :)
Random DNA data, or densely-encoded data that is effectively random at the symbol level, will be very likely to have a stop codon very early in the sequence: https://en.wikipedia.org/wiki/Stop_codon Deliberately constructed sequences can have stop codons added on purpose, although it would be extreme paranoia, since AFAIK there isn't a pathway for DNA to just wander into a cell and get incorporated, as that would be an extreme hazard to the cell. As it is they put a lot of work into making it a lot more difficult than that, and you need something at least as sophisticated as a virus to get in.
For stop codons to be relevant the first step would have to be to transcribe the synthetic DNA into RNA, which would require the presence of adequate promoter sequences. Given their complexity and spacing, this would already be quite unlikely — so most of that DNA would never produce RNA. But actual gene expression isn’t the only side-effect of introducing random sequences (although it would be by far the most severe): The synthetic sequences might also contain spurious regulatory motifs, or destroy the spacing of existing regulatory regions (not to mention that altogether random insertions would potentially destroy existing coding sequences).

At any rate, this is all purely academic since it makes very little sense to inject synthetic data DNA in living cells: Living cells have a very xenophobic attitude towards foreign DNA (as they see it as potentially viral) and try to destroy it. Once that’s solved, the synthetic data DNA now sits in a cell. Wet, and full of enzymes. These are actually not very good conditions for DNA molecules to persist. Cells age and die, and fragment their own DNA in the process. Or if they don’t, they copy the DNA and, while doing so, introduce errors.

No; DNA as a storage medium for data makes most sense when it is dried up and frozen. Better yet, vitrified (but so far there’s no way of undoing DNA vitrification without damaging it.

What'd be safer and more effective is transforming bacteria (e.g. E. coli ), which is basically an exponential photocopier.

Bacteria can be frozen down in glycerol stocks for ages, and even if the bacteria themselves are not viable, the DNA is recoverable.

To generate an appreciable amount of the synthesised DNA, this is likely already an intermediary step.

Bacteria copy DNA efficiently, but not very faithfully. As it’s not affected by selective pressure, the data would thus degrade extremely quickly. Rather than freezing bacteria, dry and freeze the DNA directly. Much better for preservation, and also much more space-efficient.
Totally agree. I was just trying to make the point that if you want to chose a living organism to stick DNA in, it's likey bacteria, not Human via virus.

Lyophilization > Bacteria Host >> Human Host

“River of Gods” by Ian MacDonald (narrated by Jonathan Keeble), one of my favorite audiobooks in recent years, was my first exposure to this idea.
Next steps are: write all internet contents into DNA spirals and engineer organism carrying them. The long term goal is to provide an access from brain to this information somehow. Hopefully it will be possible.
Shades of Johnny Mnemonic.
Next Steps : Engineer a virus to mine bitcoin using the brain power.
Then it would just become CheapCoin, wouldn't it?
A startup called Catalog Technologies has been already demonstrated DNA storage. For their demo they encoded Hitchhiker's Guide To The Galaxy in DNA.

https://youtu.be/ESGVJBXR4KE

Very cool. Long road to get to commercialization. Can you imagine a 2.5" HDD that has some crazy DNA reading mechanism in it?!
Good article however can't say that for all the articles on Cosmos Magazine. Think many articles are highly speculative and border on Pseudo Science.
Maybe one day they'll remove the latent virus code in our DNA and use those areas up for DNA Storage. Imagine if every human that went into space to colonize, had a library of information in their finger tips.
Question: with latent virus code removed, does the cell use less energy to replicate or is able to replicate faster?
In theory it should, and substantially, since DNA replication is a highly energy dependent process. In practice we can’t just nilly-willy remove genetic material from our chromosomes, even if it ostensibly has no useful function. That said, scientists have attempted long-range deletions in mouse and some of those show no phenotype (within the highly limited study!). But as far as I know they didn’t look at energy expenditure.