14
$\begingroup$

Could it be possible to drastically improve DNA repair capabilities by introducing error-correcting codes into the human genome? Would it be feasible to reduce the probability of cancer while slightly increasing lifespan?

Of course, the process of decoding a copy of one strand of DNA before transcripting DNA to mRNA would also need to be added through genetic engineering.

$\endgroup$
12
  • 1
    $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Commented Apr 2 at 17:09
  • 2
    $\begingroup$ “Error correcting codes” could apply to hundreds of very different biological processes, and most come down to “this thing miraculously happens without consequence.” Everything has a consequence. I need more specificity so I can tell you what yours is. $\endgroup$
    – DWKraus
    Commented Apr 2 at 17:17
  • 3
    $\begingroup$ We have some already, more than mice do, and not as much as E.Coli. What we have is sophisticated beyond my scant understanding, but it seems like a tough question to answer without specifics of how what you're asking about would work. To me the answer would seem to be, yes - if your specific mechanism does that then it can do that. Damned if I can see how it would work though. $\endgroup$ Commented Apr 2 at 20:30
  • 3
    $\begingroup$ Note that this still wouldn't entirely eliminate cancer. The genome might be entirely intact, with the problem being that the portions responsible for regulating cell replication are simply disabled. $\endgroup$ Commented Apr 2 at 22:05
  • 6
    $\begingroup$ Note that any error correction mechanism will be coded as DNA too, so should be resistant to mutations. And if a 100% effective error correction mechanism could be devised, it would basically eliminate the evolution of completely new traits altogether, leaving only the possibility of remixing already existing traits via sexual reproduction. $\endgroup$
    – zovits
    Commented Apr 3 at 8:28

9 Answers 9

16
$\begingroup$

If you're interested in genetic modifcations to suppress cancer, you might want to have a look at elephants and whales. Due to their large mass (and therefore high cell count) they're at a much higher statistical risk for cancer-causing mutations. In spite of this, they are able to live long healthy lives with lower cancer rates than humans.

There have been a number of studies to try to figure out why large mammals don't immediately get cancer and die. For elephants, one puzzle piece has been found already: they have extra copies of the tumor-suppressing gene TP53. Humans have this gene too, and when it's functioning correctly, it does its job rather well. 50-60% of tumors in humans involve a mutation in TP53 1. So Elephants have about 20 copies of TP53 instead of 1. If you're genetically modifying humans to increase cancer resistance, that sounds like an easy first step.

With whales we don't know as much, but there's certainly potential for in-world research that finds new cancer suppressing genes there too. That seems like a much more plausible approach than using man-made information theory (fun as that would be!).

$\endgroup$
5
  • 2
    $\begingroup$ This is mostly due to cancer turning and attacking other cancer at a certain size. So whales and elephants get cancer - all the time, but it eats itself if aggressive and denatures into binding tissue or enclosed benign cancerous masses (contained similar to a cyst). $\endgroup$
    – Pica
    Commented Apr 3 at 17:06
  • 3
    $\begingroup$ @pica those sound like interesting mechanism. Do you have e any links handy? $\endgroup$
    – N Brouwer
    Commented Apr 3 at 22:10
  • 2
    $\begingroup$ @NBrouwer: They're called hypertumors. Basically, the cancer gets cancer. $\endgroup$
    – Joshua
    Commented Apr 4 at 18:54
  • $\begingroup$ @joshua thanks - never heard of them before. Looks like they're isn't much empirical evidence for then yet but strong reasons from theoretical and mathematical models to hypothesize they exist! $\endgroup$
    – N Brouwer
    Commented Apr 4 at 21:24
  • $\begingroup$ @NBrouwer: There is empirical evidence they exist. There's no empirical evidence as to whether or not that's an effective anti-cancer mechanism. $\endgroup$
    – Joshua
    Commented Apr 5 at 0:30
24
$\begingroup$

Cellular biology already has error-correcting mechanisms that are far more sophisticated than simple codes like checksums or parity bits.

For instance, one of the most common replication errors is the so-called "frameshift error". DNA codes for proteins in three-base "codons" that are read off by the cellular machinery that processes it. In a frameshift error, this machinery attaches to the wrong place or miscounts, and so instead of a sequence of, say, [ATG][ATG][A...] it reads [...A][TGA][TGA]. Naively this seems like it should be capable of generating a huge variety of junk proteins, but in practice this is not that common. The danger of having long junk proteins leads to a selection pressure in favor of genes that are resistant to frameshift errors. Either they have "hidden" stop codons (which tell the cell to stop reading the DNA strand there) which are only encountered when frameshifted and lead to short, meaningless proteins that are easily recycled, or the frameshifted proteins are close enough analogs of the intended protein that it's not hazardous to have them around.

So, the short answer to your question is: over millions of years, the low-hanging fruit of the genome has been pretty thoroughly picked over. You'd need to come up with something quite sophisticated to outdo the systems life already has.

$\endgroup$
16
  • 4
    $\begingroup$ Frame challenge: Nature never needed humans to live past the age of 50 so there were little evolutionary pressure to resist cancer at later age. $\endgroup$
    – alamar
    Commented Apr 2 at 21:54
  • 1
    $\begingroup$ One extreamly simple way to strengthen this as an error detection code, but probably too inefficient to have evolved: Use 4 bases per codon, rather than 3. And have most of the codons code for stops, such that any frameshift of baseswap would result in a stop codon much earlier. $\endgroup$ Commented Apr 3 at 1:24
  • 2
    $\begingroup$ This answer ignores the level of sophistication of state-of-the-art error correction codes (e.g. Huffman codes and Reed-Solomon codes). They are vastly superior to things like parity bits in terms or error correction capabilities even though they are already several decades old. $\endgroup$
    – Dreamer
    Commented Apr 3 at 11:32
  • 1
    $\begingroup$ @Dreamer That's not the fundamental difference. The problem is the software or document you're downloading over Wi-Fi or LTE is needs to be bit-exact, or it will fail. Biological systems are very much the opposite: they operate with a ton of noise and highly unreliably, and rely on concepts like massive molecular-level parallelism to operate correctly. Evolution means billions of people with faulty genes will get cancer and drop dead. $\endgroup$
    – user71659
    Commented Apr 3 at 21:07
  • 3
    $\begingroup$ @Dreamer Also, I think you mean Hamming codes, not Huffman coding. Huffman coding is a data compression algorithm, and provides no error detection or correction at all. $\endgroup$ Commented Apr 4 at 1:41
3
$\begingroup$

The great thing about the double helix is how it can be replicated. Take the strand, spit it, put the complementary acid opposite the nold one to get a new double helix.

That's a simplification, of course. But the mechanism would get even more complicated if the proteins involved in the duplication had to do complex maths on top of that. For a fictional biology, you could describe Terran biochemists being puzzled and amazed by the complexity of those cells. A few crackpots might even claim that the cells are artificially constructed nanomachines, not the result of natural evolution.

$\endgroup$
5
  • $\begingroup$ I think you can just dupe DNAs without decode error-correction code. Reproduction would be quite complicated though. $\endgroup$ Commented Apr 2 at 17:38
  • 1
    $\begingroup$ Error correcting codes do not necessarily need complex math, otherwise they would be unusable for computer memory. (Because whatever math is needed has to be done very very fast.) A little bit of slightly non-trivial math is required for space-efficient error correcting codes, but if storage space is not an issue then simply storing each codon in triplicate would easily allow detecting and correcting all single-base errors. $\endgroup$
    – AlexP
    Commented Apr 2 at 17:46
  • $\begingroup$ I think I remember a story of an alien with a triple helix, such that three parts have to agree in complement for reliable replication. That would give extra error correction. Of course it wouldn't be DNA anymore as it would need different chemicals. $\endgroup$
    – Brianorca
    Commented Apr 2 at 18:03
  • $\begingroup$ The mechanism for duping DNA and reproduction are relatively unrelated. The DNA strands are not mixed at all in fertilization but the earlier process of meiosis in the parent when the gametes are formed. As such the duping process could be significantly more complicated, whilst retaining standard sexual reproduction. $\endgroup$ Commented Apr 3 at 1:41
  • $\begingroup$ @user1937198, I'm talking about DNA replication during cell division. $\endgroup$
    – o.m.
    Commented Apr 3 at 4:12
2
$\begingroup$

Different species of animals have dramatically different cancer rates even when they are in the same environment. As explained in one scientific study:

50 to 90% of aged mice die of cancer, while in humans this number is approximately 23%. Less is known about cancer in wild animals. However, several species are known to be extremely cancer resistant. These include the naked mole rat, blind mole rat, elephant and bowhead whale. The age of onset of cancer also varies greatly depending on the lifespan of the species.

This is driven by some aspect of their DNA, although not necessary crude error correction codes. For example, according to the same study:

Rangarajan and colleagues . . . showed that two hits are needed for transformation of mouse fibroblasts, namely inactivation of either Trp53 or Rb1 and an activating mutation in Hras, while five hits are needed to transform human fibroblasts (inactivation of TP53, RB1, protein phosphatase 2A (PP2A), and constitutive activation of telomerase and HRAS). Although tumors more frequently arise in epithelial cells rather than fibroblasts, this analysis suggests that humans have evolved much more robust anticancer defenses than mice.

In principle, the same DNA features that lead to low cancer rates in animals that have low cancer rates would also reduce cancer rates in humans. As another answer notes, adding "extra copies of the tumor-suppressing gene TP53" is one promising possibility along these lines.

The problem is that (from the same source): "Evolutionary pressure to evolve efficient anticancer mechanisms is very strong."

So, if an animal hasn't evolved anticancer mechanisms to a greater extent, despite this strong evolutionary pressure for hundreds of thousands or millions of years, this is probably because the anticancer mechanisms would interfere with some other fitness enhancing trait in that animal. Otherwise, all animals would have evolved traits leading to the very low cancer rates that are currently seen in only a handful of animals.

Until the genetics and biochemistry of living things, in general, and humans, in particular, are better understood, tweaking the genome to provide better cancer protection poses a hard to quantify risk that doing so will lead to other, unknown problems.

Of course, a less ethical fictional genetic engineer could simply "f- around and find out", and use informed trial and error to figure out what works best to reduce cancer incidents in future genetically modified humans at an acceptable cost in fitness reducing side effects (some of which might not even be fitness reducing in a modern environment, even though they were fitness reducing when cancer preventing traits evolved in humans and their ancestors), knowing that many subjects will suffer serious fitness reducing side effects as a result of the genetic modification.

Many notable advances in medicine (for example, the discovery of an effective way to treat fistulas by Dr. J. Marion Sims) are derived from this kind of unethical treatment of human subjects in medical research. The claim, which some historians dispute, is as follows:

Vesicovaginal fistula was a catastrophic complication of childbirth among 19th century American women. The first consistently successful operation for this condition was developed by Dr J Marion Sims, an Alabama surgeon who carried out a series of experimental operations on black slave women between 1845 and 1849. Numerous modern authors have attacked Sims's medical ethics, arguing that he manipulated the institution of slavery to perform ethically unacceptable human experiments on powerless, unconsenting women.

Using genetic modification in a crude trial and error approach to try to identify cancer resistance traits could easily give rise to a similar scandalous violation of medical ethics.

$\endgroup$
1
  • 1
    $\begingroup$ Evolution doesn't care about the lifespan of the individual. There is strong pressure to avoid cancers that inhibit the ability to reproduce. Cancers which do not would be subject only to indirect selection pressure as evolution tries to suppress the ones which do. It might be interesting to see if there's a correlation between cancer suppressing adaptations and length of reproductive viability. $\endgroup$
    – Perkins
    Commented Apr 8 at 17:13
2
$\begingroup$

All life does this to a greater or lesser extend. Fungi do it to a much greater extent than most animals so that they can safely pump cytoplasm around between cells -- which is how they can pop up relatively massive structures like mushrooms in a matter of hours.

So, yes, you can add more error correction to a genome. It will cut down on any varieties of cancer that are caused by mis-copied or otherwise damaged DNA during replication, and it will give much greater resistance to viral infection.

The downside is that it won't do anything for any varieties of cancer that are caused by genetically normal cells operating in the wrong "mode" and misbehaving, and the more you lock down your genome, the slower your adaptation rate will be, and the narrower your reproductive compatibility will be. (Of course, if you have genetic engineering and are this good at it, you probably don't care about the last two.)

It probably won't help lifespan all that much. A lot of the decisions about growth and development and repair of a body seem to be timer-based. As time goes by the timers go out of range and the ability to make repairs wanes. But you have good genetic engineering, so adding in some reset points where the body regrows various organs from scratch could do the trick. Or possibly regrows the whole thing except the brain in some kind of molting process.

$\endgroup$
4
  • $\begingroup$ Isn't a big part of the "ceases to repair itself" the existence of telomeres which limit the amount that a cell can reproduce? Those are, as I understand, a anti-cancer "measure", if some other system keeps cancer at bay, they wouldn't need this mechanism anyway. $\endgroup$ Commented Apr 4 at 10:28
  • $\begingroup$ @htmlcoderexe It's not the existence of the telomeres, it's that they wear out. You can think of them kind of like the leader on an old filmstrip. Every time you load the film into the projector you have some chance of damaging the end as you thread it in. So we manufacture them with some blank leader so you can just trim it off when you mangle it. But eventually your trimming will bite into the actual film and then you start losing video. Unfortunately, there's also some indication that telomere length gets used as a crude timer, so just preventing them from shortening may have side-effects. $\endgroup$
    – Perkins
    Commented Apr 4 at 16:16
  • $\begingroup$ Like the preamble in ethernet frames, got it - so this is actually their functionality and the "shortening as a replication limit" is a useful(?) side-effect of this? $\endgroup$ Commented Apr 9 at 6:59
  • 1
    $\begingroup$ @htmlcoderexe Pretty much. Useful because a lead-in track for DNA replication is absolutely necessary, and an organism-wide clock is highly useful for synchronizing growth and development. But, at the same time, rather a dodgy hack that's only acceptable because evolution doesn't care about the lifespan of the individual, only reproductive fitness. $\endgroup$
    – Perkins
    Commented Apr 23 at 18:40
1
$\begingroup$

This is not going to be enough. Even the most sophisticated error-correction system is going to be ultimately overwhelmed by the sheer amount of carcinogenic junk our cells get filled with over time.

It would be kinda like trying to use code correction on a script while the hardware that the script is running on is gunked with dust and garbage.

To significantly lower the risk of cancer, and to meaningfully increase lifespan, you need to find a way to efficiently clean up the cellular gunk, free radicals and metabolic trash that the mitochondria generate. Otherwise, the error-fixing mechanism is going to be ultimately overwhelmed, or the whole cell is going to crap out, and you cannot produce more cells infinitely.

$\endgroup$
4
  • 2
    $\begingroup$ This answer could be improved by going more into detail about what "overwhelmed by the sheer amount of carcinogenic junk" means. In principle, correctly configured error correction codes could handle arbitrarily large numbers of carcinogenic mutations. The limiting factor is rather that cancer could disable the error correction mechanism itself even while the error correction data stays present. This answer hints at that issue, but does not clearly explain it. $\endgroup$
    – Dreamer
    Commented Apr 3 at 11:35
  • $\begingroup$ I don't understand the premise that it is fundamentally impossible for an error correcting system to keep up. If the error correcting system is not sophisticated enough, just make it more sophisticated and more redundant. Even if the system really is inevitably doomed to fail at some point, so long as that time point is generally beyond the human lifespan, it doesn't matter. You can't truly eliminate the chance of errors, but you can make the likelihood arbitrarily low with an arbitrarily good ECC, and make the average time-to-error arbitrarily long. $\endgroup$ Commented Apr 3 at 16:13
  • $\begingroup$ @Dreamer, Im not versed enough i cellular chemistry to explain it well. Based on several articles on Astral Codex, all of which seem well researched, it seems that the ultimate reason for both cancer and cell senescence is the increased amount of Reactive Oxygen Species that come from suboptimal glycolisis. These ROS then tend to damage pretty much every cellular system in one way or another, including causing mutogenesis of all processes, even the error-correction process itself. Even if the Error-correction was super efficient, at some point the whole code would be incorrectible hash. $\endgroup$ Commented Apr 4 at 7:06
  • $\begingroup$ @NuclearHoagie Then I think we need to agree what counts as effective gain in lifespan. 10 years? 50years? How arbitrary good we want the ECC to be; scientifically plausible or plot magic? Even if we somehow made the ECC impossibly perfect, but ROS-related senescence not prevented, this will not lead to cellular life extension, but something more akin to zombiefication. You would get cell lines that divide with zero error, but are effectively unable to function, and need to get their energy from plot magic. $\endgroup$ Commented Apr 4 at 7:20
1
$\begingroup$

As both the other answers currently points out, you will have to work hard to outperform the current solution.

But it is possible. It is, however, very deep future tech. We do not know enough today to even begin to tinker with this system.

Protein synthesis is a very fundamental to how cells work and adding error correction would mean redesigning it almost from scratch. It would cost more energy to run.

Still, it is something that far future humanity might do.

An easier fix would be to find and eliminate those genes that increase cancer risk. It is well known that cancer risk runs in families. Make that stop and we will have won decades of extra life.

$\endgroup$
1
  • $\begingroup$ See also: Lazarus Long. ;) $\endgroup$
    – Perkins
    Commented Apr 8 at 17:13
0
$\begingroup$

And to introduce error checks does not fix things automatically. You just know there is a error detection and correction encoding. Like this: https://en.wikipedia.org/wiki/Reed%E2%80%93Solomon_error_correction It allows, for a space tradeoff, to restore damaged parts.

But it can not be done at the moment. The reason being is, that DNA is the ultimate spaghetti code. As in - its full of "junk" DNA, which is not useless, we just don't get what it is used for. Remove the junk, clean it up- and you get something that is not a human. Means- the shape of the useless part, like a packaging, is helping the cell heap called human, to take the shape it does. Means when https://en.wikipedia.org/wiki/Cellular_differentiation is happening, the organs and shapes do not only develop due to hormonal triggers and dna/rna expressions, but also due to the whole messy systems rolling the missed steak around.

For this to work out, you would have to do a total rewrite of the biological system called human. If you did something like this with https://en.wikipedia.org/wiki/Synthetic_biology from the ground up, you could add that as a self-repair mechanism. But we are still very far away from that.

Whats could be relatively fast possible is- a duplicate reader. Basically a virus, gets injected into a cell, where it makes a DNA copy. And it proofreads everytime DNA is transcribed - and if the proofread fails it just clogs up the cell and thus the whole thing dies.

$\endgroup$
1
  • 1
    $\begingroup$ There are some mechanisms that do basically that already built into the cell. Obviously they don't have an entire backup copy of the DNA, but they do try to kill the cell when there are too many transcription errors. One of the key things that leads to cancer is these mechanisms being suppressed. Nicotine, for example, increases cancer rates not by being a carcinogen itself, but by suppressing the cellular apoptosis reaction that tries to eliminate cancer cells. $\endgroup$
    – Perkins
    Commented Apr 8 at 17:28
-1
$\begingroup$

No.

The genome is so minuscule, and so extremely charged with electric potential, and so engrossed in surface vibration that there are no machines available to edit it. To create such a machine would require such lathing or various cutting and grasping/holding that it would require in itself that minuscule of process. Answer: No.

As an example: Consider the most advanced Central Processing Unit CPU that is being manufactured today. If you look at the lines of it connecting the transistors and at the transistors themselves via an electron microscope you observe a scene similar to a line painted by a falling-down drunk. And these are the most advanced delicate machining processes in the world. If these are so sloppy and they are the most accurate, while being so distorted, then editing a genome is far beyond that in complexity.

No.

$\endgroup$
16
  • 1
    $\begingroup$ Did you really just dismiss the entire field of nanotechnology, because it's basically "too small?" No. $\endgroup$
    – Xen2050
    Commented Apr 4 at 11:05
  • $\begingroup$ Of course there is molecular machinery that can operate on a genome, that's exactly what occurs when DNA gets unzipped and replicated in normal cellular division. Genome editing has been possible for over 50 years. What you claim is impossible has been done for decades and has seen major advances in recent years - look up CRISPR. How tiny we can make transistors has very little to do with how one edits a genome. By the logic used here, most molecular chemistry should also be impossible, simply because molecules are smaller than transistors, but that clearly isn't the case. $\endgroup$ Commented Apr 4 at 13:14
  • $\begingroup$ @Xen2050, Thank you for noticing. Almost exactly. Yes I did almost. You read that almost right. I did just dismiss the manufacturing processes required to edit at nano sizing because it is basically "too small." You almost got that correct. I did not dismiss cutting at nano sizing which can be done via laser. I did not dismiss disassembly at atomic sizing which can be done via chemical interactions. I did dismiss manufacturing at that size level. You almost got it correct. Thank you. $\endgroup$
    – Line Item
    Commented Apr 4 at 20:24
  • $\begingroup$ @Nuclear Hoagie, From sciencedirect.com [sciencedirect.com/science/article/abs/pii/… Nowhere in there is any description of manufacturing. It describes CRISPR as being a natural biological process, not a manufacturing process. Thank you for supporting my statements. $\endgroup$
    – Line Item
    Commented Apr 4 at 20:37
  • $\begingroup$ Did you read that link? CRISPR uses natural processes for genome editing. "Notwithstanding the natural immune role of CRISPR–Cas systems in prokaryotes, the most visible and significant recent research has arguably been centered around the repurposing of the Cas machinery for genome editing and transcriptional control in eukaryotes. Indeed, repurposing of the programmable Cas9 endonuclease has revolutionized genome editing and open new avenues for transcriptional control with unprecedented ease and flexibility." $\endgroup$ Commented Apr 5 at 12:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .