Junk DNA might be really, really useful for biocomputing

[extherian]

This article suggests that redundant 'junk' DNA in our genome could be repurposed as a DNA-based biocomputer.

[Tardigrada]

Biocomputing is an interesting field, it's a shame that there's not more attention to it.

[Vitto]

It has passed quite a lot of time since I studied DNA base constructs but technically, with the right sequence, you can fold it into almost any nanostrucutre.
The problem with biocomputing compared to silicon based transistor, IIRC, is more on the speed of the signals among components (which could be made of mRNA, proteins, peptides, small molecules, twisting of the same DNA) which is governed by diffusion and thus way slower than the electric field in a processor.
This if you can solve the assembling and manteinance of a DNA based chip. IIRC our 3 billion bases long genome breaks 1 million times a day and the machinery to repair it is not trivial. And for breaking I mean breaking of the chemical bonds in the double strand, then you have that pesky thing that is oxidative stress to cope with...
You also have a lot of other problems like the input-output trasducers, contamination, mutations in the DNA, etc...

Still is indeed true that the structure of the DNA and the genome allow for very fine and complex cascade signaling sequences: a single DNA gene can be influenced by enancher, silencers, promoters, DNA methilation, micro-RNA and RNA silencing in general, etc...

Biocomputing is an interesting field, it's a shame that there's not more attention to it.

From my point of view, at the moment, is already a small miracle we have a tool like the CRISPR-Cas9 to edit DNA in general with a very high fidelity (was a real pleasure sweating like a pig for the test on the techniques to manipulate DNA and having the CRISPR disclosed a few months later ).

Pretty sure we will have commercial quantum computer before biocomputing, we really still can't manipulate the DNA perfectly and even if our capabilities of sequencing has gone through the roof in the last years the understanding of how the genes works has not increased at the same rate.

Last but not least: the "junk" term is quite misleading as the DNA in these sequences can be quite relevant for the cell life and was already disregarded since I graduated a few years ago. Pretty a shame reading it in a 2020 article

[Bear]

I would say yes DNA molecules are definitely really useful for computing, but it would probably be best if we don't use them in situ in living organisms whose welfare we care about.

Junk DNA - even if it's never expressed and never affects the expression of anything - is usually valuable in some degree. Much of It is what it is, and where it is, because at some point in the evolutionary past some version of it served our ancestors well. We got a switch here, a switch there, this sequence activated, that sequence deactivated, and now it's "junk." But next week or a century from now, we get a switch there, a switch here, this sequence deactivated, that sequence activated, and the trait - or some related trait, or some part of that trait, or some combination of it with current traits - shows up in the current environment. Junk DNA is DNA recombining across time rather than just across gametes. If you go rewriting it, using it to compute, etc, you are disrupting the process of recombination across time, and suddenly tiny little mutations that might otherwise have reactivated something potentially useful or combined it with something we now have making a new useful thing, become lethal.

[Vitto]

because at some point in the evolutionary past some version of it served our ancestors well.

Plus viruses integrating in it, transposons and repeating sequences overflowing.

If you go rewriting it, using it to compute, etc, you are disrupting the process of recombination across time, and suddenly tiny little mutations that might otherwise have reactivated something potentially useful or combined it with something we now have making a new useful thing, become lethal.

I guess that "computing" here can go from activating a very simple, very localized scavenging gene package in response to a toxine entering the cell or binding to a receptor to the membrane, to "true" biocomputing, in the OA sense. In the latter case you need a knowledge of the orgainsm nearly perfect, the former is quite similar to what we do now to select recombinanting bacteria, but it would be madness do it on a genome as complex as the human one right now. Too many things that can go wrong.

[Bear]

because at some point in the evolutionary past some version of it served our ancestors well.

Plus viruses integrating in it, transposons and repeating sequences overflowing.

Well, in the sense of beings whose DNA we inherit.... those viruses are in fact among our ancestors.

[Jagace]

Last but not least: the "junk" term is quite misleading as the DNA in these sequences can be quite relevant for the cell life and was already disregarded since I graduated a few years ago. Pretty a shame reading it in a 2020 article

It is true that non-coding DNA and non-functional DNA are not synonymous, but there are portions of our genome (and large ones at that) which lack any physiological function. Most of our genome, like more than half, is ultimately of viral origin. Their sequences are tremendously repetitive, and often mutated to the point that they can't even self-replicate anymore.

That "junk DNA" doesn't exist is simply a myth and inconsistent with the current scientific evidence. An important piece of evidence is the C-value paradox: there is little correlation between genome size and organismal complexity. If you assume that all DNA is functional, then you've got to explain why a human needs 3.2 Gbp, a Vizcacha rat 8 Gbp, a bird 1 Gbp, an oak 0.5 Gbp and a species of pine 120 Gbp. Moreover, we find species that are extremely closely related and physiologically identical, such as a pair of pufferfish species, that differ tremendously in genome size.

Lore-wise, I can imagine nearbaselines and superiors simplifying their genomes, getting rid of the junky portions or repurposing them for some kind of biocomputing. For example, I could see them re-engineering chromosome 21 to fuse it with other chrosomes. The issue is that chrosome 21 is gene-poor ("junky"), so it can be duplicated to lead to viable embryo, albeit not a healthy one, causing Down's syndrome. Other chromosomes, for the most part, cause embryo death at trisomies. I could also see them doing the same to chromosomes 13 and 18, which cause truly horrific conditions at trisomies (I suggest you don't look them up  ).

[Drashner1]

In terms of RL, I can see some form of biocomputing plausibly becoming a factor at some point in the future if a number of different variables all shake out in a particular way. However, it is likely that said variables could shake out in any of many different other ways such that biocomputing never gets off the ground outside of lab experiments, whether due to being supplanted by some other tech that can match or outperform it or issues of cost and/or difficulty of integration with pre-existing infrastructure or whatever. YMMV.

In terms of OA, presumably somewhere fairly early in the timeline, Terragens figured out the 'algorithms of life' such that they can generate high accuracy computer simulations of what the results will be from adjusting DNA in an arbitrarily large number of ways and such. This would (I assume as a layperson) revolutionize the fields of genetic therapy and gengineering, and avoid many/most of the various morality and ethics issues that tend to crop up around current methods.

Just some thoughts,

Todd