Hacker News Reader

Comment by RyEgswuCsn | Hacker News Reader

It's a legacy codebase. The default assumption should be "DO NOT DELETE THIS".

Jokes aside, I wonder if we actually know enough to safely meddle with genetic materials, e.g. with mRNAs vaccines.

One problem with mRNA is that 75% of its name is what it's built out of, and only 25% what it does. It'd be like if we called HTTP packets "hBits", flash memory words "mBits", text files "tBits", and OS kernel instructions "kBits". I can see how someone would be worried that putting tBits on their computer could potentially modify their kBits! What if all those bits got all jangled up!?

If we called mRNA a "protein print job" like pjc50 came up with, and we called our chromosomal DNA our "kernel instructions", etc., I think it'd be more obvious how silly it is to worry that mRNA is going to somehow alter your chromosomal DNA. The point is that there are a lot of different functional things and data structures made out of bits, and a lot of different functional things and data structures made out of adenosine, guanine, cytosine, and thymine/uracil.

To be fair (not trying to imply mRNA vaccines modify your DNA), there are a lot of systems, especially past but also present (with vulnerabilities), where the right (wrong) print job could end up permanently modifying your system's kernel instructions.

Watch out, your computer is using electricity that could potentially kill you.

mRNA vaccines don't modify your dNA. They provide a volatile blueprint of the corona-specific spike protein, so that spike proteins can be synthesized for a short time, analyzed and then discarded by the body.

mRNA vaccines don't meddle with your DNA. They inject mRNA in to cells that _temporarily_ instruct the cell to produce a specific protein. Once the mRNA is exhausted, the cell stops producing that protein.

So in that case the mRNA is more like printer ink than like a blueprint? Or both at the same time? Is it possible for a cell to still produce the protein when it runs out of mRNA?

I'm not an expert in this but... If DNA is like a blueprint, mRNA is like an email sent from the architect to the contractor to instruct them to carry out specific work.

It's a print _job_. The cell is printing more bits of cell all the time, we just slip in some extra messages that print foreign proteins. The immune system identifies the proteins as foreign and produces antibodies to destroy them, which primes it to destroy matching COVID viruses.

As an ex-biologist, this is the best analogy.

mRNA is "messaging" RNA. It sends the message to the protein factory to make more of a protein structure defined by the DNA, just like a print job sends the message to the printer.

If DNA is a blueprint, mRNA is a photocopy that's used as a working memory of it for a ribosome to use while making a protein. It's chemically unstable and soon falls apart. (The individual nucleotides night be recycled into other mRNA strands that the cells make.) Since the vaccine mRNA does not have any corresponding DNA to make more copies from, that's the end of it.

> It's chemically unstable and soon falls apart.

Thanks, that was the part that the others answers didn't cover.

The "ink" would be amino acids that are always floating around in the cell, the "printer" would be a ribosome.

The mRNA would be something like the print buffer that stores the currently printing document. DNA would be the document file stored on disk.

mRNA vaccines are barely meddling with genetic materials. We know that lots of viruses are barely harmful; mRNA vaccines are basically just:

• take out most of the steps

• change the payload

• substitute one of the ingredients to trick the immune system.

If you said CRISPR, I'd agree – but we do know enough for mRNA vaccines.

To use a CS analogy, mRNA vaccines aren't Turing-complete, but CRISPR is. We understand both fairly well, but the implications of mRNA are much simpler than the vast, vast, vast implications of CRISPR.

I'm curious why you question CRISPR. I've worked with CRISPR; I've seen some very unexpected results from its use; I've also not really kept up with the science in the past few years.

Do you have experience working with CRISPR? Is there a study or set of studies that concerns you, or is it the reality that negative results don't get really get published?

There's nothing in the comment that "questions" CRISPR, it's just pointing out that due to the larger space of possibilities more care needs to be taken with it.

I interpreted the comments about CRISPR in the parent comment to be more or less cautionary, and I used "question" as a way to say "not sure about the benefits". Sorry if I was unclear, I will work harder on my communication.

CRISPR is basically find-and-replace with DNA. DNA expression is weird and complicated, and some bits do different things depending on the environmental context. There's a lot of potential for things to go wrong.

mRNA vaccines are injecting simple, engineered mRNA. This mRNA is expressed 100% of the time; what it does is very simple. I'd be confident using a new mRNA vaccine on day 1, so long as they're certain they picked the right protein, but I wouldn't be confident using a CRISPR treatment without a full medical trial.

I know what CRISPR is.

Like I said, I worked with it in grad-school. I agree with you in general, though I will say that if I was in a position to get CRISPR as a treatment, I would probably have a lot of questions for my doctor, but I wouldn't dismiss it outright.

I guess my point is that our understanding of how mRNA vaccines operate (or any other methods for manipulating genetic materials, for that matter) might prove incomplete, just like how we underestimated the importance of microchromosomes, as stated in the article.

The current generation of mRNA vaccines may be simple enough for the scientists to analyse, but I feel it will only be a matter of time before the complexity of such technologies grow to the point that it resembles art more than science. (I am looking at you, deep neural networks!)

mRNA's mechanism of action is trivially simple. There's a known gene that codes for a known protein, and they create a snippet of mRNA to codify it. It doesn't change your DNA, it's not a virus, it's not able to replicate in any way. There aren't any unknowns waiting to be discovered. The mechanism is fully-understood and easy and safe.

The complications with mRNA vaccines like this have to do with keeping the mRNA intact during delivery and getting it into the cell so the protein gets built. That's the hard part, and the focus of all the patented engineering that differentiates BioNTech and Moderna vaccines. And the delivery mechanism, while complicated, doesn't involve anything genetic; it's a mechanical problem.

> There aren't any unknowns waiting to be discovered. The mechanism is fully-understood and easy and safe.

I am sure many people felt that we had figured it all out w.r.t. planet motions etc. with Newton's law of physics :)

You are mistaken if you think I am anti-vaccine or something. I merely tried to point out that there are always unknown unknowns and that we should remain cautious when it is people's health that are at risk.

CRISPR is not "turing complete", I'm not sure where you got this impression. You could build a pseudo-turing machine (finite memory, measurable but small error rate) with enzymes and DNA, but CRISPR is not sufficient to implement turing complete computation.

He did not try to say it was turing complete. He was making an analogy: CRISPR:mRNA::Turing Complete:Non-Turing Complete. Not sure how deep that analogy is, since non-turing complete systems can still be very powerful. But I think it gits the idea across OK.

OK, it was written very misleadingly. Further, the comparison doesn't even make sense. RNA molecules can fold into ribozymes and carry out activity. In principle, you could make a turing machine with RNA molecules, although the error correciton would be fantastically hard.

Computer scientists, please stop making CS/biology analogies.

But that's not how analogies work.

When someone says "A is to B as X is to Y", you're not supposed to start comparing B to X. Your post is comparing B to X.

Let me try to come up with another example... Let's say I make an analogy like "a house relates to a wall the same way a forest relates to a tree". (It's not a very good analogy, but that doesn't matter here.) If you look at that analogy and then start talking about how you can use a tree to build a wall, you are doing analogies wrong. The items on the left of the analogy are not supposed to be compared to the items on the right of the analogy. You're only supposed to compare the relation on the left to the relation on the right. Your discussion of RNA-based Turing machines is doing analogies wrong in the same manner. Turing machines are on one side of the analogy and RNA is on the opposite side.

That's also not what they wrote, they used it as an analogy - in the sense that mRNA can express one very specific thing, and CRISPR can express a much wider variety of things. It's not about literal turing-completeness.

Reply on news.ycombinator.com