I expected to have to scroll through pages upon pages of indecipherable text. Instead it's no bigger than a large paragraph of text, and I can easily fit it on my screen.
The technically challenging parts are:
- delivery mechanism: you need to take a very unstable molecule, protect it from the environment - both external, and when inside the patient - and insert it into a human cell. (This is called the "platform", and is usually developed independently from the specific payload.)
- manufacturing: both producing the mRNA itself at a large scale, and inserting it into the delivery mechanism, at a large scale and in low-temperature conditions
- testing: the newly-developed payload and the existing platform were integrated at small scales within weeks, but testing the thing for safety and efficacy took months
EDIT: As schoen pointed out, this was not actually released by Moderna, but reverse engineered by third-party researchers. Original text was: "Hence they feel safe releasing this. Their moat is not the gene sequence, their moat is everything else."
One or more of the vaccine developers may have released such details, but this particular file is a reverse engineering effort by unaffiliated scientists based on analyzing the dregs of used vaccine vials (!).
Edit: See https://news.ycombinator.com/item?id=26628594 for more substantive discussion about this.
From the CDC "ingredients" for the BioNTech/Pfizer vaccine, along with cholesterol (which modulates the stability of lipid membranes), they report using this molecule, which would form a phospholipid bylayer, just like our own cells use: https://www.sigmaaldrich.com/catalog/product/avanti/850365P?...
Of note, the immune system is pretty good at destroying foreign mRNA so you also need to evade it.
This article is pretty good: https://berthub.eu/articles/posts/reverse-engineering-source...
(Though RNA may have been more stable under the high-UV-exposure conditions the early Earth.)
https://www.modernatx.com/sites/default/files/US10702600.pdf
though they do present multiple sequences, so I guess you'd have to go to the FDA application to figure out exactly which one got used.
I have a "I'm sure that means something to somebody" feeling. It's also surprising that the remaining claims seem to describe the resulting bits of the sequence, and that that primary claim can stand on its own. Of course, I'm by no means an expert.
But a specific composition that encodes its spike protein, encapsulated in a lipid nanoparticle? That's much more of a creative work.
So somewhat like how a fishing fly differs from the insect it represents.
Probably worth noting that patents are not required to be creative. That's copyright terminology.
I think the terminology used for patents is something like "inventive and useful".
I think that's the question. I'm very much used to reading "A machine-readable medium, comprising..." I'm curious what bits are unique to COVID-19, and what bits are generally protecting the idea of using a carrier to send a specific protein. In this case, is it the "S" protein phrasing that protects the specific embodiment of COVID19's spike?
Aha! That's what is surprising to me - I assumed that that would have been done previously / protected previously. That explains it.
Break it down! It's not so bad:
> A composition
A bunch of stuff
> a messenger ribonucleic acid (mRNA)
mRNA are cellular instructions on how to make proteins that are read by ribosomes that make those proteins as they read along.
> an open reading frame
This is something that starts with a "start codon" and ends with an "ending codon" and encodes valid instructions to make a protein between.
> encoding a betacoronavirus (BetaCoV) S protein or S protein subunit
The instructions refer to the spike protein of a betacoronavirus, or a fragment thereof, because this is what we want the immune system to pay attention to (and make antibodies to bind to and neutralize).
> in a lipid nanoparticle
The immune system gets pissed off about mRNA floating around, because that's one of the things that happens with active infection. So if you want this to get into cells and tell them to make your protein, you need to encase it so that it mostly escapes immune notice itself.
https://twitter.com/vesselofspirit/status/123748864242514739...
According to the song lyric, "novel coronavirus / has a lipid outer shell". So it seems like some viruses have taken advantage of this.
Anyway, I think the important thing that the other commenter was saying is that mRNA needs to be carefully packaged to be medically useful. You can't inject pure RNA as a vaccine because not only is the mRNA going to be quickly degraded before it gets anywhere but if the immune system sees any RNA floating around by then it kicks itself into a frenzy because free-floating RNA is usually a sign that something nasty is afoot.
Moderna vaccine has to be kept in -70 C. Viruses that can survive for any extended period of time only in -70 C won't find many hosts.
Also maybe the process of creating virus shell can't naturally be done without protein scaffold.
The envelopes used in an RNA vaccine are generally simpler, because they're working under different constraints than viruses. For example, their envelopes don't need to be easily manufactured in a cell.
But some RNA and DNA vaccines do use viruses as their delivery mechanisms, eg the J&J COVID vaccine.
While this particular Moderna claim would likely affect BioNTech/Pfizer's mRNA vaccine, it's not clear whether it would survive in litigation, too.
As to a "specific string"-- if you could just pad a few codons onto the end and not be violating the patent, that's not too worthwhile.
Animation of transcription from mRNA: https://youtu.be/TfYf_rPWUdY
Transcriptase: https://youtu.be/5MfSYnItYvg DNA polymerase: https://www.youtube.com/watch?v=bee6PWUgPo8 The Ribosome: https://youtu.be/TfYf_rPWUdY
The mRNA vaccines work in much the same way-- it's just that the mRNA vaccines only include the code for the spike protein and not the rest of the virus's machinery. So you get the vaccine and your body produces a bunch of spike protein by itself, which gives your immune system the opportunity to learn how to identify and react to the spike protein before it sees it on a real virus.
meh, I could do that over a weekend never sounded so scary, or impressive at the same time. That weekend just so happened to stand on the shoulders of prior decades of research though.
i guess this is big pharma's version of `apt-get install`
I just don't think it was every profitable enough for them to put in this enormous capital expenditure.
As in, "do not generate enough income"? Really? Now?
We don't really need the same quantities of that, though.
It can be said that hepatitis care is more necessary than covid.
That translates to a total cure cost of:
75M * $350k = $26.25T
There are other ARV treatments available which cost now roughly $50-100k and cure in 3 months.
Whereas immunizing everyone against coronaviruses currently costs:
8B * $7 = $56B
Clearly, the costs of the HCV cure are predatory and unreasonable because it doesn't lead to eradication and it's inaccessible to the poor and the third-world.
The true unit cost of HCV cures is unknowable but possibly half of the current price.
For all the horror HIV has wrought, global spending on vaccine development for HIV has been around $1 billion a year for the last few years. In contrast, the USA federal government spends $3 billion a year for HIV antiviral drugs for low-income Americans. $20,000 per patient per year for life. Unsurprisingly, new antivirals are where most of the research is.
It can sound almost like a conspiracy if I put it like that, but it's just the economics incentives. Especially since the developed countries where most of the market for charging a decent markup is, have the smallest market for most new vaccines, while having the largest markets for therapeutics for chronic conditions.
Now HIV is genuinely devillish to develop a vaccine for despite our attempts. But vaccines for hepatitis C, gonorrhea, HSV, among others appear to be possible. We almost certainly could develop effective vaccines for these with existing techniques, if someone coughed up the funding. Maybe all the buzz about mRNA vaccines will spur some progress here.
But since the demand is 14 000 000 000 doses, there should still be a little bit of money in it?
The parent poster was describing the situation with vaccine development in general, to which COVID-19 is quite the exception. A potential hepatitis C vaccine for example has very different economics, as it would not be deployed anywhere nearly as widely or quickly. Consider that, 40 years after hepatitis B immunization became available, the majority of Americans haven't been jabbed with it.
Talk about market failures! It's completely obvious that this economical system is not placing the good of the whole human species as its first priority.
Sounds like a problem you solve once and for all, for any vaccine. And also that this problem was already solved since decades (e.g viral vectors)
- testing: the newly-developed payload and the existing platform were integrated at small scales within weeks, but testing the thing for safety and efficacy took months And so many people have been killed by this overly conservative testing, phase ~<2.5 was enough
I strongly doubt it. It's more like a problem you solve once for a particular class of payload and particular destination. Biology doesn't do packet switching - everything is just rapidly bumping into everything else at random, so your envelope needs to be designed in a way that's ignored by everything else than molecules at your target site, and it needs to not react with the payload it's carrying.
> And so many people have been killed by this overly conservative testing, phase ~<2.5 was enough
Overly conservative? That's what super-accelerated testing looks like. We're lucky it went well; had they screwed up, it would scare a lot more people away from vaccinating, lengthening the pandemic and increasing death toll.
What kind of tweaks were made from "the version they threw together in a weekend" to "the version that is in production now"? What's a typical "mRNA" feedback iteration loop like?
After the massive capex that has gone into mass-production of encapsulated mRNA delivery systems, I suspect this new technology will be very cost-competitive for the big markets like the annual flu vaccine.
https://blogs.sciencemag.org/pipeline/archives/2021/02/02/my...
Why manufacturing of these vaccines is a hard part.
The most amazing thing is that now that the platform is proven secure in dozens of millions of people, it should be be very easy and fast to get approval for other payloads. Biontech for example wants to go after cancers - a platform that can deliver payloads targeted to an individual's cancer is nothing short of a game changer in cancer treatment because the current standard of blasting the patient's body with a lot of highly toxic chemicals is arcane compared to letting the body's immune system do the cleanup.
re: cancers, that is actually what this technology was originally developed for! Moderna has been spending about a decade getting this tested and proven out for the cancer role, and they're quite close. From my quick reading of the literature, there seems to be some regulatory confusion about how exactly to run approval for this kind of personalized drug design (testing the method of generating the individual drugs?), but the bar is usually much lower for cancers with high mortality rates.
Or the distribution method, or even really invent the thing, since you're mostly just copying someone else's work. Plus it doesn't have to even do anything. In fact, doing anything might be a problem, so best to just sit there and look menacing (and spikey).
Coincidentally, the mRNA sequences for both vaccines are about 4kb (kilobase) long.
Getting it designed and building it is more difficult.
At its core, it’s a piece of mRNA that creates a protein. That code gets transcribed into a protein (often those are relatively short). That protein then triggers your bodies immune response, which trains it to attack covid19.
Inject this mRNA into a cell and it’ll create the protein. Anything can be injected at this point once the mechanism for injection is developed
Not sure where the technology is exactly at, but I suspect we're no more than 5 years from major incident related to this.
Even this vaccine, we really don't know the long-term impacts or risks involved with this. For instance, this vaccine does appear more risky than the standard flu vaccine:
https://wonder.cdc.gov/controller/saved/D8/D137F981
Presumably this is due to increased inflammation. It's not hard to imagine that we'll be doing genetic editing soon enough with this (if we aren't already).
Also, you might not even be able to print full viruses with this platform. manufacturing mRNA is different from manufacturing all the random types of RNA in a virus, isn't it?
There are people already doing gene editing with mRNA methods:
https://m.jpost.com/health-science/israeli-scientists-use-mr...
I’m not sure about these particular platforms, but I wouldn’t be surprised if we see gene editing technology deployed live in the next 3 years.
Excerpt from the disclaimer in your source:
> VAERS accepts reports of adverse events and reactions that occur following vaccination. Healthcare providers, vaccine manufacturers, and the public can submit reports to VAERS. While very important in monitoring vaccine safety, VAERS reports alone cannot be used to determine if a vaccine caused or contributed to an adverse event or illness. The reports may contain information that is incomplete, inaccurate, coincidental, or unverifiable. Most reports to VAERS are voluntary, which means they are subject to biases. This creates specific limitations on how the data can be used scientifically. Data from VAERS reports should always be interpreted with these limitations in mind.
Wish there was more but the incentives aren’t really aligned for open research on this.
We can get the real cdc idc10 (billing) results in several months.
This should hopefully provide you with some useful perspective.
Biology is a funny old thing. You can look at that concise description - the orange and so on blocks of a few letters and a few short groupings.
Now ATCG are basic building blocks but they consist of quite a lot of stuff. I think it's a bit more complex than that because this is RNA not DNA so ATCG might not be quite right. Each of those bases are horrifically complicated depending on scale. Search "ATCG" - this is a good start: https://en.wikipedia.org/wiki/Nucleobase
Now dive into one of those bases and decompose it to its constituent atoms. Now look at the maths around this stuff. It gets quite complicated, quite quickly.
That said, the fact that a bloody complicated thingie can be described so concisely is absolutely amazing and as you say it looks so simple.
> So in the BioNTech/Pfizer vaccine, every U has been replaced by 1-methyl-3’-pseudouridylyl, denoted by Ψ. The really clever bit is that although this replacement Ψ placates (calms) our immune system, it is accepted as a normal U by relevant parts of the cell.
Neat.
"Many people have asked, could viruses also use the Ψ technique to beat our immune systems? In short, this is extremely unlikely. Life simply does not have the machinery to build 1-methyl-3’-pseudouridylyl nucleotides. Viruses rely on the machinery of life to reproduce themselves, and this facility is simply not there. The mRNA vaccines quickly degrade in the human body, and there is no possibility of the Ψ-modified RNA replicating with the Ψ still in there. “No, Really, mRNA Vaccines Are Not Going To Affect Your DNA[2]“ is also a good read."
As far as I could tell, this would work well for getting a synthetic virus into the human body, but without the necessary mechanics within our cell, the special Ψ chemical won't be reproduced by the virus. That'd mean the replicated virus would get snatched up by the immune system as soon as it'd get released from the cell.
Theoretically, a complex enough RNA string could be used to have our cells build the necessary cellular machinery to properly reproduce the virus, but that's a kind of altering DNA that's a whole different can of worms. There's probably cheaper and easier way of defeating the immune system, for example by simply "enhancing" ebola or HIV to make them more infectious and more resistant to our current drugs.
[1]: https://berthub.eu/articles/posts/reverse-engineering-source... [2]: https://www.deplatformdisease.com/blog/no-really-mrna-vaccin...
It's often the protein molecules that the immune system learns to recognise and attack.
RNA vaccines work because your body automatically translates them into some recognisable part of the viral protein, and then develops an immune reaction to that.
If a virus had Ψ instead of U in its RNA, it's still going to be making the same type of proteins. I can't see why it would be more likely to evade an immune response.
Second, our immune system doesn't just attack and recognize free floating RNA, but the virus itself. And different parts of the virus. First and foremost it will recognize the surface proteins (like the spike protein) because those are the things that it can see while the virus is outside of the cell. Also these are the things that the infected cells present on their surface (MHC II sites, if I'm not mistaken) to the immune system. (As far as I can understand, cells have to present the proteins they produce to the immune system otherwise they get killed. If they produce alien virus proteins that get recognized by the immune system, they also get killed.)
Interesting enough, the immune system somehow also recognizes the so called nucleocapsid protein, which is the one used to wrap the viral RNA inside the virus. (But it gets produced by the cells, so I guess they get presented on the cell surface so the immune system can learn to recognize and counter them.) I didn't look into the details too much, but as far as I can understand it's not clear yet how those antibodies (the ones created against this protein) work, because antibodies are supposed to be used outside of the cells, but the nucleocapsids are only present inside the cells and then inside the virus.
To sum it up: the immune system is much more complex, has several recognition mechanisms, the viral RNA is mostly packed into the viruses (or are inside the cells) and the viruses don't have any way to produce Ψ (or any of the other nucleotides).
You are correct, the antibodies are made and they end up not recognizing the nucleocapsid protein while it is in the virus but when the infected cell displays internal proteins with MHC I, which helps T-Cells target the infected cell. MHC I displays self and MHC II displays proteins that have been "eaten" by the surveillance cells of the immune system.
2) Your immune system does not usually attack the RNA housed inside a virus, but rather protein fixtures on its "body".
"Since each of the 20 amino acids is chemically distinct and each can, in principle, occur at any position in a protein chain, there are 20 × 20 × 20 × 20 = 160,000 different possible polypeptide chains four amino acids long, or 20n different possible polypeptide chains n amino acids long. For a typical protein length of about 300 amino acids, more than 10^390 (20^300) different polypeptide chains could theoretically be made. This is such an enormous number that to produce just one molecule of each kind would require many more atoms than exist in the universe."
I mean, I can understand how an eye or a brain can evolve by natural selection, but I’m still stunned by abiogenesis. I guess we’ll never know for sure how it all started.
The whole thing is absolutely fascinating and wild.
Honestly, na. It's pretty verbose. There's a lot of weird ass things in there like "Skip basepairs until you find the matching terminating sequence" (I think it's AG .* GA but its been a decade since my bioinformatics course), but you still have to include the non-AA-coding basepairs in the middle of that.
Compensating for that is the fact that there are like, multiple independent programs; if a ribosome is offset by a single base pair, the result is entirely different. If it runs the other strand, the result is different. And instead of crashing like any program would, biology just learns to use all of those possible encodings. In part, this works because there are 64 possible codons but only 20 amino acids, and the redundancy allows a substitution to affect only some of the offsets.
...with GATACCA right in the middle, but unfortunately with no GATTACA that I could find.
For comparison, the smallest chain that they technically call a protein is 100 amino acids that's an arbitrary limit to separate proteins from enzymes. So this thing isn't tiny tiny.
But Titin (also called connectin), a giant protein responsible for passive elasticity in mucles, is ~27,000-35,000 amino acids. So this thing isn't even close to the biggest proteins out there.
Do you mean “to separate polypeptides from proteins”? Enzymatic activity has nothing to do with size. For example, one of the smallest enzymes in humans has 62 amino acid residues. And, under certain conditions, even single amino acids can be catalytic.
But yeah, the polypeptide-protein threshold can get fuzzy, especially with the recent advances in miniprotein characterization.
The story I remember was that Insulin was the first protein that was sequenced, which is funny because it was before they made the distinction. It's actually too small to be considered a protein now.
When I saw it, I thought that it could almost fit in a tweet, so I just did it:
https://twitter.com/weinzierl/status/1376807707957719041?s=2...
The sequence takes 16 tweets, 15 if you don't split at line endings and remove spaces (4175 nucleobases / 280 nucleobases/tweet ~ 14.9 tweets).
Remind me the joke of the consultant engineer knows where to make X by the chalk. LOL
[1] https://berthub.eu/articles/posts/reverse-engineering-source...
I don't know how long it will be before we get a bit more serious with it, but geneticists have a big obstacle in their understanding, any change might needs a thousand strong lifelong population study to be understood. That's way crappier than dumping the assembly or only having the documentation in Chinese.
I will add that moreover the developers might have been even more conservative in their code because they knew it was going for large scale deployment, they probably avoided the cutting edge as much as they could.
Lots of these things aren’t complicated. It’s the careful systematic testing and public trust building that’s the hard part.
https://www.nytimes.com/interactive/2020/04/03/science/coron...
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5584442/
There’s also (IIRC, no citation right now) prior work suggesting that coronavirus vaccines against the spike are likely to be effective and that vaccines against the N protein might be counterproductive.
Make your own, open-source. Really cool.
A user on lesswrong made their own (with no prior experience): https://www.lesswrong.com/posts/niQ3heWwF6SydhS7R/making-vac...
Only two companies in the world succeeded, the French company Sanofi which also tried making a mRNA vaccine failed.
I presume a whole bunch goes into making vaccine and this is just the top of the iceberg.