The was an article in The Atlantic a couple+ months back - which after 10 mins I still can't find - about some scientist and his alt view / theory on DNA. The gist, if I recall correctly, was (and forgive me if I fall short of capturing the significance of his idea):
How DNA functions is far more complicated than how current accepted science pictures it.
1 - Certainly seems like he might be onto something.
2 - If so, could the use of CRISPR be premature, if not dangerous?
> 2 - If so, could the use of CRISPR be premature, if not dangerous?
This is really an ill-posed question. First off, CRISPR is a tool. Its novelty is the fact that it makes an experiment vastly simpler and more reliable — but, fundamentally, the experiment was already possible beforehand. So CRISPR itself isn’t dangerous. At most, modifying genetic material could be.
Secondly, genetic modification — whether by CRISPR or other methods — is an extremely generic process. Sure, it could be dangerous, if you use it to perform dangerous experiments. Can it have unintended side-effects? Sure, we already know about those. In fact, one of the great things about CRISPR is that it drastically decreases side-effects.
But is CRISPR dangerous specifically because we underestimate the complexity of DNA? No. Nothing about CRISPR assumes that we know all biological processes that involve DNA. In fact, we know that we don’t know them all.
Lastly, somebody with an “alt view on DNA” immediately sounds like a crackpot with a poor handle on modern biology. That may sound harsh but biology is a science, not a pseudoscience or protoscience (or magic): we’re moving within some relatively fixed boundaries imposed by experimentation and evidence. Uncertainty or lack of knowledge isn’t a free-for-all to insert wild speculations that usually amount to “everything we know is wrong”. It’s very rare that “everything we know turns out to be wrong”. In reality, subsequent findings refine existing ones, or disprove specific hypotheses. They don’t disprove an established field. Even something as revolutionary as Einstein’s theory of relativity didn’t do that: gravity continued to function just fine afterwards. Newton continues to be taught in school, because Newton’s laws continue to be relevant.
> Sure, it could be dangerous, if you use it to perform dangerous experiments. Can it have unintended side-effects? Sure, we already know about those.
In my opinion, this is the wrong way to frame this. We know there are unintended side-effects, but you make it sound like we already know what we don't know. Therefore, everything is relatively safe... which can't logically be true.
You can't gauge the specific danger or benefit of the unknown.
> You can't gauge the specific danger or benefit of the unknown.
I’m not claiming to. All I’m saying is that “CRISPR [isn’t] dangerous specifically because we underestimate the complexity of DNA”.
In the context of this discussion I can go much further, though: the physiological presence of i-motifs doesn’t change the hazard or risk of CRISPR whatsoever. We know this because our risk assessment of CRISPR never relied on the non-existence of i-motifs.
Is this necessarily true though? Do you think the situation I illustrate here is not possible (crappy photoshop of cas9 guide RNA matching a DNA sequence across a 'imotif'):
No, I think it’s totally possible (albeit exceedingly unlikely). But we already know that CRISPR has (rare) off-target effects. That’s why I’m saying that the i-motifs don’t change the risk assessment. At best, they can explain why CRISPR sometimes misbehaves (although it’s much more likely due to just plain old nonspecific binding in most cases).
Parent and microbiologists can correct me if I'm wrong, but the primary drivers of current DNA tampering are correlating different forms of genes with health outcomes.
We aren't creating new genetic code ex nihilo (yet).
So understanding the nuances of how different forms of genes influence outcomes is less important than knowing that they do. And what the naturally-occurring "correct" form looks like.
One of the main drivers of CRISPR-Cas9 biotech development is to perform targeted single-nucleotide DNA editing, to cure or significantly reduce the risk of getting a disease.
Numerous diseases have a genetic component; some people have a higher risk of getting a disease due to their genetic makeup. Often a significant amount of the risk is attributed to a 'single nucleotide polymorphism' or 'SNP' (this is known to be true for diseases like sickle-cell anemia, cystic fibrosis, Alzheimer's disease, and others).
If we could alter this SNP however, from the disease-associated nucleotide to the control nucleotide, theoretically this would substantially reduce the risk of getting that disease. CRISPR-Cas9 allows us to do that. This process relies on fabricating an RNA-guide sequence to match the segment of DNA we want edited; and if we are careful, we make sure this Guide RNA matches ONLY this segment of DNA and nowhere else in the genome. If the Guide RNA matches some other random segment of DNA, it will result in 'off-target' edits. Now, this shouldn't really be an issue if we perform whole-genome sequencing on an individual prior to making the Guide RNA. We can just make sure the Guide RNA doesn't match anywhere else in their genome and all is well, right?! Well... maybe not. Maybe these i-Motifs will fuck us over like in the picture I linked above.
Not being a jerk. I'm truly interested in filling the blindspot in my "software."
That said, the gist of the guy / article was (and I'm spitballing) CRISPR is an over-simplistic view of DNA. Sure, it might work for a couple of things, but to presume that pov / lens is OSFA is likely not correct.
> "Lastly, somebody with an “alt view on DNA” immediately sounds like a crackpot with a poor handle on modern biology."
Right. Because the world was and still is flat? SMH
If I had $20 for every crackpot in the history of science I'm be FU money wealthy :) __The whole point of my comment_ was to say this new tread/artice, along with the mentioned article in The Atlantic seems to be saying it's __possible__ our understanding of our current understanding of science is off-base.
With other gene editing techniques, such as TALENs (and others before that).
> And if so, why is CRISPR such a big deal?
CRISPR/Cas9 is a big deal because it makes gene editing a lot easier and less error-prone (and hence cheaper, etc.). But (and not to diminish the tremendous scientific achievement behind the technique) the hype in the press around this technique is only partially warranted, and a lot is getting lost in translation. Part of the reason why CRISPR specifically is a big deal in the popular press is due to the extremely dirty fight for its exclusive patent (and the Nobel prize for its discovery): To gear up support for their side, both institutes involved created substantial media coverage that far exceeds what new discoveries normally get.
> CRISPR is an over-simplistic view of DNA
I would really like to read that original article because stated like this, the claim is simply incorrect.
> Right. Because the world was and still is flat? SMH
I honestly don’t know what you mean by the rest of your comment. But it’s worth noting that, in the modern history of science (i.e. since doing something that actually deserves that name), crackpots have virtually never been right. Progress (even paradigm shifts) came always from experts inside the field. The only near-counter-examples I can think of are Semmelweis (but that was before modern biology existed, and even he was an expert in his field), and Wegener (likewise, for continental drift). It’s easy to assert that science doesn’t have all the answers, and that theories are a temporary view of the world that’s going to be proved wrong in the future. But this has never been accomplished by people who doubt the veracity of extremely well established facts.
Our knowledge of DNA is far from complete. But the things we know about it today will stick around: they have been tested by millions of experiments, performed around the globe each day. Paradigm shifts will come because we will discover new things about it, not because we’ll disprove established facts.
My understanding is that CRISPR is popular because it is much simpler and cheaper than other previously available methods
"So what makes CRISPR so special?
It's much easier to employ than older gene editing technologies, and it also has a very high success rate.
Because it is more user-friendly to the average scientist, it has opened the door to all kinds of research that previously would have been too expensive or required too much time to carry out." [1]
"...biology is a science, not a pseudoscience or protoscience (or magic)"
I agree with your overall point but I take slight issue with this statement.
I am not biologically educated, but I am biological. I have no idea what my conscious is or why/how it works, even though I do know that I exist. And neither do you, nor anyone.
To me, that is the classic definition of magic. This differs from eg electronics, physics etc, because I am biological. I have a personal perspective on it, in a manner that I could never with electronics, physics etc. Ie. I can readily understand how far away from understanding the biological process humanity is.
The layers don’t stop because you choose one: you are biological, because you’re chemical, because you’re physical (which is more like saying you’re electrical).
Ie. Laws of physics govern the laws of chemistry and the first couple are chiseled in stone for the most part. Biology, that is to physics what psychology is to surgery (perhaps mathematics is a batter analogy).
You can’t choose that you are ‘just’ biological because it is the emergent property of deeper systems of rules.
Consciousness, now that is a debate for the philosophers.
> This differs from eg electronics, physics etc, because I am biological
biology has everything to do with physics and electronics (and chemistry etc).
it’s just a frame of reference that decides what you choose to classify as one or the other. But the laws of nature are consistent no matter what you’re choosing to call it or which aspect of it you’re addressing.
Knowing about the existence of unintended side-effects is not the same as predicting their frequency. Ultimately the reduction of risk comes from hedging against an expected frequency of risk, something that we are still learning how to do, or else CRISPR would already be a widespread mode of therapeutic treatment.
It seems what you're actually trying to argue against is becoming bearish on new technology due to risk, but you're doing this with posturing that is ungrounded in both the details of the biology and the risk.
> or else CRISPR would already be a widespread mode of therapeutic treatment
You underestimate how long it takes for anything to get from fundamental research to treatment. Not just because of risk. And CRISPR is already in extremely widespread use as a molecular biology tool.
Apologies for going off tangent but, as someone who majored in biochemistry/molecular biology (now almost 8 years in the past, which was pres CRISPR but has been reinforced by continual education in medicine) This is such a beautiful inadequacy of our knowledge.
My intro to science and molecular science was first massively sparked by an E. coli article I randomly jumped to in the early stages of Wikipedia. It talked about error correcting proteins, and versions of this and that, and I thought holy shit the cell is a computer.
As I came to understand more of molecular biology I was fortunate to find it very easy to visualise and understand not just the idea that the theories and molecular pathways we understand are an amazing set of dominos that have to fall in a particular order but also the stochastic relationship of them all - on, off, according to molecular affinities, and those are just the pathways we have been able to understand by way of intelligent experiments and human ingenuity.
To jump radically back on topic -
There have to be and are so many more layers to how the code that builds life executes that we don’t understand that it is pure arrogance to assume that we understand the system enough to perfectly execute any modifications to it. That doesn’t mean it’s wrong to look down the rabbit hole, after all, that’s all we’ve ever done
The The Atlantic article is sensationalized and oversimplified. The scientists are mainstream, and they found that most of the variation between people in complex traits such as height are due to variation in genes outside core pathways. In other words, there is relatively little variation in genes responsible for growth hormone; instead the cumulative effects of millions of variations in seemingly unrelated genes is what explains differences in height.
This is not relevant in the context of gene therapy, where we only change one gene or a small number of genes. From that paper, we would expect that a single change would have a tiny, negligible effect on unrelated genes.
Moreover, the initial targets for CRISPR therapies aim to correct mutations. For example, cystic fibrosis is caused by a point mutation in the gene for a mucus protein. CRISPR would be used to restore the genotype that all healthy people have. Since healthy people have this gene, side effects are unlikely. (However, we do expect side effects from other mechanisms such as immune rejection.)
The danger of using CRISPR could have, in principle, been worse than it turned out to be, sure. But the fact is, CRISPR has been used to modify living organisms, and those organisms don't have any more mutations than average, and they seem to be just as healthy as unmodified organisms (unless the edit itself is deleterious).
We might find out some day that there's some behavior of DNA and DNA-binding proteins that CRISPR interferes with, but if it does exist, the effect has to be small enough that we wouldn't notice it.
edit: I should clarify, there's still work left to do to ensure CRISPR's safety. It needs to be looked at in more cell types, and work in somatic cells in adult organisms is still in its early days. My point is just that putting CRISPR into a cell, in some scenarios, doesn't cause substantial problems.
> How DNA functions is far more complicated than how current accepted science pictures it.
So, specifically: i-motif structures were understood experimentally. They had never been detected in vivo before, and doing so constitutes great science and a solid find, but no: this isn't something out of left field that our naive hubristic minds couldn't conceive. These scientists found i-motif in vivo because they were actively looking for it.
Beyond that, what you say is true of course, but only in a specious sense that applies to every field of science. The whole point to doing these experiments is that we don't understand things completely and want to. I mean, would you argue against air flight in 1903 based on the fact that a computational framework for dealing with the Navier-Stokes equations was still decades in the future?
> How DNA functions is far more complicated than how current accepted science pictures it.
From what I learned in courses on edX (e.g. MITs Eric Lander's "Introduction to Biology" or RiceX BIOC300.2x "DNA: Biology's Genetic Code"), that is already well known and taught. Scientists are not naive and overconfident in their knowledge, individual examples to the contrary notwithstanding. Guess what - they teach what is actually reliably and verifiably known. It's just a bit too much to expect scientists to work with methods and knowledge they don't even know exist yet, or even to teach them - but that doesn't mean they think they already know it all.
Science is full of this exact thing: Layer upon layer of models. Bohr's atom model is very far from what a physics student learns these days. The Newton universe vs. the relativity universe. Chemical bonds in high-school, then in early university chemistry courses, and then what physicists work with who study bonding seems like it comes from entirely different universes, although it all describes the same thing.
Same with DNA. The more courses you take the more the knowledge resembles a Mandelbrot set image. Actually, that's a horrible analogy, because in that image when you go deeper you still get similar structures, but when you go deeper in science the same thing starts looking very different. The reason why I still use this analogy is because the part that seems applicable and important is that you can go deeper and deeper and deeper and... (etc.) However, as far as I could tell from just a few courses that complexity already is being taught as far as this is possible (as I mentioned above, it's hard to teach "unknown unknowns", to quote a famous American politician philosopher /s ).
What you should keep in mind though is that on each level of knowledge there are a lot - often A LOT - of experiments and applications of that level of knowledge that have been shown to work reliably and well. It's just that new levels of knowledge (and models) open up new possibilities. "Every model is wrong, some models are useful", I only understood what that means when I learned more and more and realized how useless the word "truth" is when attempted to be used in absolute terms. Each model is "true" for the things it successfully and reliably predicts, and for the applications it makes possible. Finding deeper models does not invalidate what you already learned and used successfully.
I would suggest seeing science not as the search for "the one final truth". As practised, it is far more practical (yes, same word twice, deliberately). You find something, you demonstrate that it works, you show that it is useful ("you" is many people), you move on. "If you want 'truth' you are in the wrong lecture, philosophy is down the hall" is a sentence I remember to have read having been said a (biology? physics?) professor when he started the first freshman lecture. By concentrating on what you can demonstrate you don't need to have "the final truth" before you are able to do anything useful. You go step by step, and while you don't know if there ever is a final word that does not really matter. That means your statement
> How DNA functions is far more complicated than how current accepted science pictures it.
actually already is built into science. You can always assume that there is much more that you are not yet aware of.
I wonder if the formation of an i-motif could change the behavior of the "normal" DNA code. Like a flag on a function to slightly change how it works.
Could make certain parts of reading normal double helix DNA a no-op, for example by accumulating parts before they can be read.
Another idea is maybe this is just a structure for temporary memory for use in a more complex "algorithm".
We used to think DNA was just mostly static data but now it seems there is some computation going on.
If not computation to change behavior, maybe the i-motif act like a check sum to make sure a complex or essential part of DNA isn't corrupted (by a storage failure or a virus)
I dont think this motif brings anything spooky/special to the fore. DNA is swarmed by proteins, ribocomplexes, acetyl/methyl histone mods, etc. all of its own making. Whatever special computation one might speculate this motif could perform- that computation or fidelity check or whatever else is probably a known and well-documented function of some DNA-associating complex. While it's fun to speculate how this might 'change the game', in all likelihood, this finding is a very small drop in a very large bucket.
But if we are speculating...
That a motif forms by a single strand of DNA or RNA is expected (given certain properties or symmetries). What the image in the article depicts and given the quote that this is often seen at start codons, to me, suggests this motif forms whenever transcription machinery has opened the helix creating two single strands. And instead of hanging out as single strands (a very vulnerable state, to DNase degradation, etc) they coil unto themselves. This might also suggest this segment of DNA is relatively easy to unzip, and has built-in degredation protection, making for ideal consensus sequences as start codons.
Seems like these "i-motifs" form transiently during transcription.
If they are making the argument these motifs exist much longer than the course of transcription I'd like to how DNA could remain stable in such a conformation. There are molecular machines inside nuclei that really dont like single stranded DNA, and will chomp it up on sight. If the complementary strand to the i-motif isn't itself moonlighting as some double stranded motif, how is it protected from these dnases.
If they are only saying the conformation exists while DNA is undergoing transcription... this isnt exactly on par with a A-,B-,Z-DNA conformations.
The primary source is behind a paywal for me (while off campus), but I'll be interested to read more about this later:
So it seems like there are two timescales for what 'transient' means in the paper.
1) There are different levels/numbers of i-motifs identified depending on the cell cycle position (highest at G1/S boundary, although this is only comparing cells synchronized to G1, G1/S and early S so maybe more at different points?), suggesting these structures don't just stably form and then just sit happily around for the entire lifetime of the cell.
2) In vitro these motifs are less stable than (say) G-quadruplexes so presumably there is a suggestion they may be transient over short timescales, but this is not actually examined in the paper. No idea how you'd actually test this without inherently perturbing the equilibrium being examined in vivo. Even if you could avoid fixing the cells antibodies would be out of the question because of the inherent linkage (if you bind the i-motif with a 500 pM Kd [high affinity] you're gonna HUGELY stabilize that conformation).
"What excited us most is that we could see the green spots – the i-motifs – appearing and disappearing over time, so we know that they are forming, dissolving and forming again,”
anyone know what technique they used during this observation? It sounds.. not fixed?
what matters most in my opinion is whether this is a spontaneous DNA conformation or whether this is a transient conformation DNA assumes while transcription machinery is preparing to read nearby bases. (if the former, kinda interesting; if the latter, much less interesting)
Right - yeah; I don't know, I had the same thought when I read the commentary (I read the paper first). Nothing in the papers suggests the ability to monitor formation/loss in realtime I don't think? My guess is this is an interpretation of the data from the fact that you see lower levels in G1, higher levels in G1/S and lower levels in early S - i.e. they must be 'transient' because the levels go up and down again.
Seeing this happen in real cells in realtime would be – I would have thought – technically almost impossible. We're at the cusp of viewing the formation/loss of clusters of RNA POL or mediator clusters with the most advanced super-res (see Ibrahim Cissé's work) but these are comparatively massive protein clusters, so the idea of being able to view DNA structural transitions at [effectively] single-molecule resolution where that transition involves a few nucleotides in a non-perturbative way seems like a reach.
Seems like the obvious next step is to break 'em with synonymous mutations and ask if there's any detectable phenotype.
"No idea how you'd actually test this without inherently perturbing the equilibrium being examined in vivo. Even if you could avoid fixing the cells antibodies would be out of the question because of the inherent linkage (if you bind the i-motif with a 500 pM Kd [high affinity] you're gonna HUGELY stabilize that conformation)."
I mean, I trust the reviewers/editors of Nature but I don't understand how this is not a serious confound!
Those of you who are CS people who are trying to stretch this paper into more significant than it is: please go and read all the first-year biology textbooks before speculating.
I am a CS person. If the article is not readily consumable for a CS person why is it on a website called "Hacker News"? Seems like a bad submission if you ask me...
> I am a CS person. If the article is not readily consumable for a CS person why is it on a website called "Hacker News"? Seems like a bad submission if you ask me...
Like many before you, you appear to assume Hacker equates to the C in CS.
Like the many that will, after me, assume Hacker to mean the C in CS. Technically, you are correct, and I agree. But I must point out, that actually, you are wrong - the general use of the word belongs to computers.
Most people would have no idea about Richard Stallman, and I maintain that due to popular media the general public associate the word with computer technology.
I'm a biohacker (by training) but a computer hacker by vocation. It's clear the HN community has an abiding passion in the deep relationship between information and computation systems. I just want to make sure the less experienced readers don't immediately jump to making naive conclusions by making weak analogies between computers and biology.
In the Guidelines link to at the bottom of this page under 'What to Submit':
On-Topic: Anything that good hackers would find interesting. That includes more than hacking and startups. If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual curiosity.
Off-Topic: Most stories about politics, or crime, or sports, unless they're evidence of some interesting new phenomenon. Videos of pratfalls or disasters, or cute animal pictures. If they'd cover it on TV news, it's probably off-topic.
The problem right now is that "has only been witnessed in vitro – that is, under artificial conditions in the laboratory, and not inside cells." which means it may not even exist in our bodies.
But this is literally the entire point of the this paper. The sentence you just pasted was the the previous state of the art, this paper makes the point that they have now imaged them in cells (using i-motif specific antibodies).
'in vitro' can mean 'in cells in a dish'. I think the previous state of the art was 'unsure if these exist in living cells' (could be wrong, but id be surprised if they did this exp in vivo). edit based on your comment above you probably have the paper and are in a better position to comment on this. I defer to op.
Yeah agreed (what in vitro means to different people is a whole other conversation :-P), but that sentence makes it pretty clear that by in vitro they mean not in cells (because it says, "and not inside cells"). Agreed, of course, that this doesn't necessarily mean it happens in cells in an organism (though in the authors' defense they do examine three different cell lines).
Edit: For clarity - all the in cell work is cultured cell lines and not cells taken from an animal model or in situ imaging.
Literally my only point was that 'this might not matter for biology because we've not even seen it in cells' is no longer true. It still might not matter for biology though!
Computation: “What excited us most is that we could see the green spots – the i-motifs – appearing and disappearing over time, so we know that they are forming, dissolving and forming again,” says Dr Mahdi Zeraati, whose research underpins the study’s findings.
Hooray! Finally the inevitable happened! I guess this will turn biology upside down, everybody assumed DNA/RNA were static (with the exception of some primitive viruses).
Another example would be the discovery that the ECM (extracellular matrix) plays a huge role in cellular development. Completely unrecognized until experiments done by Mina Bissel's lab opened up an entirely new field in cellular biology.
Superresolution microscopy. It blew through a long-standing limitation- the best resolution was determined by the wavelength of light passing through a lens. That and other recent discoveries in microscopy like light-sheet microscopy have turned microscopy on its head.
No, I'm agreeing with you. I think it rare that any mature field these days is "turned on its head". I doubt most of the examples given by other comments would qualify if you dig into them.
A good example would be induced pluripotent stem cells nobody thought that that can happen but once it was obvious that you just needed a few transcription factors to move cells from state-to-state. Completely trans formed a field that was already mature
I mean by that a model of computation in the TCS sense. (Some) molecules, be them DNA, RNA or other, are akin to lambda terms (or graphs), beta reduction (or other graph rewrites) are akin to chemical reactions. The algorithm of reduction has to be random, local (i.e. no higher point of view criteria for interaction are used), performed by Nature. All in all, to be able to prove that some part of the biochemistry can be seen, as a concrete asynchronous graph rewrite automaton (capable of universal computation) would be a great advance in biology.
I'm not sure why that would be the case. First, I think most smart people already recognize that sufficiently complex biochemical systems can in fact perform computation as well as other abstract operations. Second, there is already a ton of literature in this area, it's often very helpful to analyze biological data from the perspective of various graph and sequence rewrite automata, but that doesn't necessarily mean that the underlying biology is explicitly exploiting matter at that level. Third, I don't see why having any such "proof" (to the extent that you can prove anything about biological systems anyway) would be an advance. What new predictions would you be able make?
The DNA is changing real-time, i.e. there is a periodic process we usually call algorithm. Algorithm might imply some sort of computation going on; while it's pretty common to see computations being performed via neurons and protein interactions, seeing it directly in DNA is much more interesting. Instead of assuming genes are static, there actually could be some DNA-level computation going on, potentially invalidating significant portions of our knowledge of genetics.
We already knew this. DNA is a highly dynamic molecule, and it’s changing constantly, as it’s involved in biochemical processes in the cell. i-motifs aren’t special or new in this sense.
Full text PDF: https://paperpile.com/app/p/ce9c364d-416a-0ede-9233-adf31910...