AlphaFold Found Possible Psychedelics
nature.com
nature.com
But the really important conclusion of the paper, which I truly hope bears out, is that AF2 models are useful for finding binders to GPCRs.
GPCRs are one of the largest (if not largest) drug targets and work (including my own) showed that modelling protein dynamics using expensive simulations was likely to be necessary to make accurate predictions of GPCR binding molecules. If this paper is correct, AF2 is actually able to recapitulate much of the detail that we previously believed required expensive dynamics simulations, which is about the best possible news because it suggests that we can omit these costly calculations.
It's almost certain now that AF2 will win the Nobel Prize, as this represents some of the most exciting breakthroughs in the past few decades.
The 3d structure of GPCRs is intimately associated with the cell membrane, with one end sticking out into the extracellular environment, and another sticking out into the interior of the cell. Simply getting that sort of protein, which is not soluble in water, was a massive challenge and the first structures didn't arrive until the early 2000s, when people learned how to crystallize GPCRs using detergents and other solvents that simulated cell membranes.
So many diseases are caused by misregulation of signal trandsduction... let's take some examples. Vasopressin is a drug that also happens to be a natural body product. In some diseases, people don't make enough vasopressin to regular the kidneys properly, leading to a form of diabetes, and simply giving people more of it helps reduce the symptoms. I guess in this case (not certain), people have mutations in the GPCR that receives the vasopressin signal that attenuates the signal "too much" and by just dosing the patient with a lot more, more signal gets sent inside the cell.
I think at least 10 nobel prizes in medicine have been awarded for research related to GPCRs, which has uncovered a wide range of medically relevant knowledge of physiology. The NERSC supercomputer "Cori" was named for Gertie Cori, who helped discover some of the core mechanisms in human metabolism during the golden age of metabolic biochemistry.
I think the simplest way to think about it is being able to target specific GPCRs is like being in front of a switchboard with buttons and knobs controlling every critical detail of how human bodies manage and maintain themselves.
And as the article says, >"There were a lot of models that we didn’t even try because we thought they were so bad,...”
With all the associated bullshit and snake oil and romanticizing of the "good old natty days" too, of course.
Big $$$ industry, at least until our brave future does away with voluntary reproduction altogether, being the reason people bother to exercise in the first place.
I do hope people choose a nice walk over color match games.
Agree finding new 5HT2A ligands is cool, what would be truly interesting is having a suite of totally selective ligands for the entire 5-HT receptor family.
I feel like a lot psychopharmacological research into a lot of mental disorder is severely hampered by
1) the complex interactions of the various serotonin receptors(and also with other receptors like D2)
2) the lack of truly selective ligands, necessitating a hodge podge of less selective agonists combined with selective antagonists to isolate the effects of specific receptor interactions(making it hard to do this sort of thing in humans). Or using knockout mice, which is also kind of dubious because you've now made a fundamentally different brain and are trying to compare it to a normal one. And that's sort of hit and miss.
What seems more likely to you: that the (obviously worse [1]) structures produced by alphafold are somehow magically capturing protein dynamics, or…the virtual screening method (DOCK) used by the paper is relatively insensitive to the quality of the underlying model? They say right there in the article that neither screen was obviously enriched versus the other (the hit rate was around 50% for both branches, which was essentially the same as for the very different retrospective screening method in the same paper [2]).
I’m putting my money on the parsimonious answer. Running structures through a large enough virtual screen will almost always find something that binds, even if the starting model isn't great.
[1] at least in the case of the x-ray structure. Using a cryoEM structure is sort of interesting, since they’re blurry anyway.
[2] the actual paper: https://www.biorxiv.org/content/10.1101/2023.12.20.572662v1....
I think you're right, that running lots of structures through virtual screens will find lots of binders. In the middle days of protein design (well before AF2), people went from not being able to design anything, to being able to easily design "rocks"- proteins that were exceptionally stable, but didn't do anything interesting. Their loss function that they maximized for was stability. But that comes at a cost: high stability seems to prevent lability (which is useful for enzymes). And I think that if you just pass enough stuff through a good-enough program, you'll find lots of binders, but those binders won't actually be useful binders. We already know from decades of work that we can find lots of nanomolar and picomolar binders that don't do anything worthwhile (or have side effects).
My conclusion is: we may very well be on the tip of solving another bottleneck in the drug discovery pipeline, or, we might just be bumbling around in a big space and found a nice local minimum. I don't think AF2 or similar systems on their own are going to completely solve the problem, and until DeepMind or somebody else creates something that can truly address the hard problems in human biology, the most we can hope for is finding more effective new drugs more cheaply and somehow couple that with improvements in genetic personalization.
It seems weird to give the peace prize to a model. What about the people who made the model, or is that what you mean?
Unsure of the merits of this finding myself, but I'd be cautious given just this past week DeepMind was accused of publishing gibberish by Chemists: https://twitter.com/Robert_Palgrave/status/17443839652705816...
I agree the chemistry paper does not look g ood but it's hard to say yet. Like I've said elsewhere DM and Google Researchers are under huge pressure to make their results look as good as possible. Already people have serious concerns about another Google paper using ML for IC layout.
What happens when that labor isn't as valuable? It seems like someone in power would either want to pacify us or eliminate us.
https://en.wikipedia.org/wiki/PiHKAL
> PiHKAL: A Chemical Love Story is a book by Dr. Alexander Shulgin and Ann Shulgin, published in 1991. The subject of the work is psychoactive phenethylamine chemical derivatives, notably those that act as psychedelics and/or empathogen-entactogens. The main title, PiHKAL, is an acronym that stands for "Phenethylamines I Have Known and Loved."
> The book is arranged into two parts, the first part being a fictionalized autobiography of the couple and the second part describing 179 different psychedelic compounds (most of which Shulgin discovered himself), including detailed synthesis instructions, bioassays, dosages, and other commentary.
https://www.vice.com/en/article/ppzgk9/interview-with-ketami...
Its apparently almost as strong as LSD, takes hours to start working and the effect lasts several days. Worse in every respect. Why would anyone ever choose it?
and also, the stories in Phikal and Tikal of more disciplined people following titration protocols etc. and discovering hidden kingdoms of the mind, is the stuff of dreams.
Very few people are going to want to try RCs if conventional alternatives don't present more legal risk. Most people using them are avoiding legal risk or trying to beat drug tests etc.
There is a tiny minority of "true psychonauts" who are in it for the experimentation, but we tend to be acutely aware of the risk and take much better precautions than someone who is chasing a high they're addicted to.
And adults knowingly taking risks(on their own behalf) and suffering consequences is just the human condition.
My recollection is inexact, but there was a story about a person who synthesized a novel theoretical psychedelic, tested it on himself, rapidly developed Parkinson's-like symptoms which deteriorated until his death. I don't know that Shulgin was ever or always the first to try novel syntheses, but I do think allowing known-safes is better that challenging people to put themselves at risk through self-experimentation.
>The neurotoxicity of MPTP was hinted at in 1976 after Barry Kidston, a 23-year-old chemistry graduate student in Maryland, US, synthesized MPPP with MPTP as a major impurity and self-injected the result. Within three days he began exhibiting symptoms of Parkinson's disease. The National Institute of Mental Health found traces of MPTP and other pethidine analogs in his lab. They tested the substances on rats, but due to rodents' tolerance for this type of neurotoxin, nothing was observed. Kidston's Parkinsonism was treated with levodopa but he died 18 months later from a cocaine overdose. Upon autopsy, Lewy bodies and destruction of dopaminergic neurons in the substantia nigra were discovered.
https://en.wikipedia.org/wiki/MPTP
Seems you were mostly right, with a few caveats:
1. The toxin was an impurity in his synthesis of MPPP, which is a synthetic opioid.
2. He was diagnosed with parkinsonism and treated, but died from a cocaine OD(one wonders whether the cocaine was an attempt at self-medication, or a deliberate suicide attempt).
According to the article this case of MPTP-induced parkinsonism has happened again in the 80s, again with tainted batches of MPPP.
I'd say this young chemist was more similar to Shulgin than most psychonauts, but he was also skipping steps and clearly had issues with addiction.
This would likely be avoided by Shulgin because he'd probably get the chemistry right in the first place.
Shulgin also had a lot of pharmacology understanding that allowed him to make educated guesses about what kind of structures might be very dangerous.
But he definitely was the first to both synthesise and ingest a lot of these more obscure psychedelic compounds. But he was extremely careful, which is why he lived well into his 80s.
How many of the AlphaFold protein folding predictions have been actually verified by a lab? And are any of those actually new proteins or just tiny variations on things in the training set?
> Shoichet agrees that AlphaFold predictions are not universally useful. “There were a lot of models that we didn’t even try because we thought they were so bad,” he says. But he estimates that in about one-third of cases, an AlphaFold structure could jump-start a project. “Compared to actually going out and getting a new structure, you could advance the project by a couple of years and that’s huge,” he says.
AI is definitely capable of solving novel programming problems and I'm constantly surprised how well Copilot understands the esoteric context of my code.
GP could've read the papers, or read commentary of the papers, or read about the 13th Critical Assessment of Structure Prediction (CASP) contest.
Instead, "llms are bad, so isn't this bad too"
I don't believe that. I think you are being fooled by the enormity of the training set, followed by the non-novelty of things you personally haven't seen before.
I'm not saying AI tools aren't useful though. I'm just saying they are what they are: a fuzzy search in an enormous problem space. This is impressive and cool, but not the same as being able to solve truly novel problems.
Us and plants aren't creating the same molecule because we share a common ancestor. There are far too many other plants that don't create cannabinoids or opiates or whatever, and far too many other animals that don't create them, for this to be at all likely. It had to have evolved multiple times.
It's not any kind of symbiotic relationship. Our ancestors didn't cause opiates to evolve by selectively eating and spreading their seeds.
So it's a coincidental convergent evolution. But these are really complex molecules doing quite different things. Both cannabinoids and opiates evolved in plants to defend them from viruses and fungi. I don't know what other neurotransmitter-mimics are for, but they're not being used for plant nerves.
I'm sure people know the answers to this, and I could read a book about it.
I don't think they do. AFAIK most of your assertions above are unknowns.
My theory, at least re: psilocybin and the like, is that a parasitic fungus developed neurotransmitter agonists as a way of manipulating its host. The complete cocktail that was needed to make the target species behave in a way that completed the fungus' lifecycle may be lost to time, but certain components of it still exist in the gene pool.
And if you're going to have millions of spores (rather than the two or three offspring that, say, humans have) you can afford to set them up with a wider array of gene expression profiles. So you've got these fungi just trying stuff to see what works--now armed with neurotransmitter agonists as part of that "stuff".
Later on you've got other organisms eating those mushrooms and behaving strangely. If you're a fungal spore, and you're in an animal's digestive tract, "strangely" is probably how you want it to be behaving. You're going to end up in a different place or with different nutrients or in some other situation which is abnormal for your host--you end up with greater habitat diversity.
How specifically that host-behavior-strangeness would lead to a lifecycle that would reinforce the inclusion of psychedelics to the point where P. Cubensis is reliably psychedelic, rather than just the occasional mutant, I have no idea, but when you've got billions of years to play with, it's not hard to accept that some pretty weird animal-fungus interactions came up here and there.
https://en.wikipedia.org/wiki/N,N-Dimethyltryptamine
https://en.wikipedia.org/wiki/Melatonin
Just a slight mutation, or even biological synthesis under non-ideal conditions could easily result in synthesizing one or the other via the same pathway.
There is a lot of complexity in a million years of mixing spores and mycelial karyotypes. I haven't done the math but it seems like it could easily rival a human brain.
I think marijuana or poppies have less plausible explanations for how they would have evolved specifically to mess with animal's minds.
Sometimes we need a little shove to start doing new things that end up being advantageous. There's no reason to believe that we're unique in that way. If that shove is advantageous enough to be worth returning for a second shove, well now you've got the seed of a cross-kingdom relationship.
That is, in fact, exactly the reason.
You are of course correct that synthesis of many biological molecules has evolved countless times over the eons, but chirality of certain critical molecules, amino acids, ribosomes, all evolved exceedingly early. Once the cellular machinery is synthesizing these things, synthesis of similar molecules, one, two, three mutations away become likely across swaths of related organisms, however distant.
In short, we're all playing with the same lego bricks, and there are only so many different models possible to build with the set.
The synthesis of opium, for instance, requires two specialized cell types, complex organelles, and hundreds of different proteins, all told requiring thousands of different genes [1]. This is all for the purpose of creating one single highly-complex molecule. This isn't something you get at easily by "playing with the same Lego blocks."
The question is, if it's not a coincidence that both poppies and a small subset of animals both go through all this work to create two versions of this extremely-specific molecule, then what's the reason?
Sure, in the end the reason will be "it's advantageous for both species," but it's still pretty mind-blowing, and I think we're still missing a bunch of steps.
Also, this is not a refutation of your point, just interesting, but complete opioid synthesis has been engineered into yeast by at least two different teams at his point: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4924617/
But like many things in biology, if you view it through a billion-year evolutionary lens, it's not completely surprising. I doubt anybody could come up with a general rule or law that explained this well, and anything I did would just speculative arm-waving, of which this site has enough.
Both sets of species are each doing a whole bunch of work to create these very specific, very complex molecules, and using them for different purposes.
Today's babies don't require psilocybin to become conscious.
Use of psychedelics could have helped accelerate an individual's "spiritual" development and therefore afforded them additional offspring. So the correlation, if any, might be a favoring of genes that paired well with the effects of psychedelics. Pure speculation on my part.
As for AI, it seems highly likely a new class of designer drugs will emerge, tailored to a specific individual to enhance societal adaptation. We've already seen this in play over the past 40+ years in the pharmaceutical industry. AI will only accelerate it IMO.
It makes you wonder how this brittle system got its start. Considering everyone today needs this language kick start induced by cultural practice, maybe the initial kick was also some cultural practice that also introduced new ideas where there wouldn’t be otherwise. Maybe that practice at one time was a shamanic like experience considering the proliferation of these practices in the most basal religions we have around the world. Its really hard to say what came first, the language or cultural practices such as ritualized drug taking, considering what archaeological evidence we have seems to suggest these things were happening at the same time.
The War on Drugs is a lot bigger hinderance than technology here. And the patent-driven pharmabusiness too.
It can cause permanent damage to vision or sanity. Sure, it's not addictive, and you practically can't overdose -- but it makes you temporarily insane and blind, and both of these can persist without there being a cure. And just, a general point: [CITATION NEEDED]. Unlike you, I'll provide my own citation: https://www.frontiersin.org/articles/10.3389/fpsyt.2017.0024...
> And there are a lot of similarly safe compounds with "smoother" psychological effects.
If you mean the party drug 2C-B, it has some similar risks.
Almost all drugs (illegal or legal) have potential side effects. Most psychotropic medicine have quite nasty side effects and/or cause dependence. SSRIs can make you insane or to lose your libido. Benzodiazepines are very addictive and their discontinuation may cause fatal delirium. Eating a box of Tylenol can easily destroy your liver permanently.
Comparably, LSD is pretty darn safe. At least when taken in well controlled doses and in safe settings. Hard to see LSD's side effects being a showstopper for its medicinal use.
Yes, 2C-B and the many many similar compounds tend to have similar side effects.
The 5HT2A receptor is comprised of 471 amino acids. Proteins fold based on interaction between each of those amino acids, plus the external environment (plus post-translational modification).
So imagine a length of string with 471 magnets attached to it with different magnetic field strength (ranging from weak to strong). Now imagine it floating in space - how would the string fold upon itself? Now imagine different segments of the string have different stiffness - some are super flexible, others are rigid.
Difficult to predict how it might fold? Oh yeah!
Now imagine the string not in space, but in a solution of water with a number of other magnets floating in it (the cytoplasm within the cell and external environment outside the cell). Oh, and also imbedded in a wall with magnets in it as well (the cell membrane).
Now how does it fold?
Ok, now we think we have a good approximation of the structure just based off of various pieces of data. But wait! Other proteins are also embedded in the cell wall - all with their own structures and selection of weak and strong magnets, stiff and flexible segments interact with the receptor! Oh, and not all receptors have the same external environment - that can differ based on another laundry list of factors. Yeah, that's another layer of complexity.
Then add on top that the external environment (inside the cell on one side, outside the cell on the other) has a ton of water, but also dissolved ions like Na+, K+, Cl-, plus a bunch of other proteins with their own "magnets" and "stiff and flexible segments". These outside molecules will bind and disassociate from the receptor on a constant basis, ever so tweaking the shape.
Oh crap.
Of course AI has a good starting point - we can get x-ray crystal structures of the receptor and identify the position of atoms. But wait! That's in a crystal, not a human body! So that structure is likely incorrect in several ways. We also have other analytical techniques that can give clues to the structure in a solution that more closely mimics what's inside the body, so we can at least tell when the x-ray structure is just wrong (at least in some places).
That should give people an idea just how complex protein folding is. AI is good at taking all known "rules" about protein folding (and molecule interactions with the protein) to improve the predictions. But we're still far off from truly understanding all the factors in play.
I thought they are all small molecules.
/s