AlphaFold developers win $3M breakthrough prizenature.com
196 points by dopu 5 days ago | 79 comments
dekhn 5 days ago [-]
I expect them to get the Nobel prize in Chemistry in about two-three weeks.
mxwsn 5 days ago [-]
The Nobel committee usually prefers to wait and evaluate longer-term impact, so I'd be quite surprised. CRISPR was obviously revolutionary in 2013 (imo, more than alphafold), and won the Nobel in 2020.
riedel 5 days ago [-]
> more than half a million researchers have used the machine-learning system, generating thousands of papers

Funny things is how the general scientific community (including nature) defines 'impact'. I somehow still strangley trust the Nobel committee to take a different approach here. Was curious and found this interesting collection of references: https://www.researchgate.net/project/Enacting-Excellency-Awa... .

echelon 5 days ago [-]
CRISPR is a revolutionary tool.

AlphaFold doesn't solve folding. It makes metaheuristic guesses without writing a bunch of quantum chemistry, statistical physics, thermodyanamics, and topology maths / algorithms.

I don't mean to downplay AlphaFold, but we haven't solved protein folding yet. This press is really getting ahead of itself.

m00x 5 days ago [-]
Are you a trained scientist, especially in the biochem/medsc field? What makes you more certified to say this isn't a revolutionary tool?

I personally know several people that do research that have unlocked new possibilities through this tool. My wife is a neuroscientist and she's used this tool a few times for reasons that are above my head (even with a Msc in Microbiology). This type of work used to take a PhD student 4 years or more to do a single relatively simple protein. Getting answers within a few seconds is revolutionary.

avsteele 2 days ago [-]
acapybara 5 days ago [-]
This reminds me of the criticisms of Copilot (GPT-3 application to generate computer code).

Many engineers will say it doesn't code. It just regurgitates and remixes the data it was trained on. It just makes "meta-heuristic guesses."

But anyone taking an honest and objective view of it can see that Copilot does add value. It's no substitute for a real, human, engineer, but it clearly adds value.

I don't think AlphaFold would get to this level of funding, resource commitment, etc. if it was adding 0 value.

Being a domain expert, I'm curious what value, if any, you think a large transformer model could add to the domain of protein folding. Is it really zero value, in your view?

misnome 5 days ago [-]
> I don't think AlphaFold would get to this level of funding, resource commitment, etc. if it was adding 0 value

They didn't say that it added zero value, that's entirely you. They said that it doesn't solve the problem on its' own, which is true.

m00x 5 days ago [-]
The computer doesn't solve problems on its own, but it is objectively a breakthrough technology. It's pretty clear that AlphaFold is a phenomenal innovation.
mnd999 5 days ago [-]
Copilot adds value to Gitlab if that’s what you mean.
falcor84 4 days ago [-]
Did you mean "GitHub", or did I miss the joke?
mnd999 4 days ago [-]
Yeah, you missed it.

GitHub ripping off everyone’s code to build copilot has caused a small (not as big as it should have been imho) exodus of open source projects to Gitlab and others.

I still don’t see how it’s a good idea to import unknown code with unknown licenses into your project, but apparently if copilot does it it’s okay.

cercatrova 5 days ago [-]
> The AI effect occurs when onlookers discount the behavior of an artificial intelligence program by arguing that it is not real intelligence.

> Author Pamela McCorduck writes: "It's part of the history of the field of artificial intelligence that every time somebody figured out how to make a computer do something—play good checkers, solve simple but relatively informal problems—there was a chorus of critics to say, 'that's not thinking'." Researcher Rodney Brooks complains: "Every time we figure out a piece of it, it stops being magical; we say, 'Oh, that's just a computation.'"

https://en.wikipedia.org/wiki/AI_effect

echelon 5 days ago [-]
Let's be precise here.

Solving physics isn't a soft thing like making pretty art.

I'm firmly in the "AI/ML will eat the world" camp, but the praises being foisted upon AlphaFold are borderline damaging to the real field and its practice.

You can't throw AlphaFold at pharmaceutical problems and call it a day. This press feels like a "mission accomplished" victory lap when it's very clear we're only just getting started.

m00x 5 days ago [-]
I have a Msc in microbio, not biochem, but my understanding is that proteins don't have a constant shape. They vibrate, interact with other molecules, etc.

You won't have a perfect answer unless you want to predict its shape in a vacuum, which wouldn't be very useful either way. Having it "close enough" is already extremely useful. There are definitely edge cases where it gets it wrong, but there are always edge cases in ML. More data = better results with the same architecture.

Tons of things that won Nobel prizes weren't 100% accurate, it's not a prize for solving science, rather a prize for advancing science.

My wife is a neurobiologist and the impact of this advance is groundbreaking for her work.

bongoman37 5 days ago [-]
frozencell 5 days ago [-]
Echelon is revolutionary too, Peace Nobel prize worthy.
SturgeonsLaw 5 days ago [-]
The same Nobel Committee that gave Obama a Peace Prize within months of taking office?
kirrent 5 days ago [-]
The Norwegian Nobel Committee which selects the recipients of the peace prize doesn't have anything to do with the physics and chemistry prizes which are awarded by the Royal Swedish Academy of Sciences.
SturgeonsLaw 5 days ago [-]
Thanks for supplying the extra detail as a counterpoint to my flippancy. I was aware that the Nobel Prize for Economics was "not a real Nobel Prize" but didn't know the Peace Prize was also quite separate from the science based awards.

It does make sense though, seeing as the scientific awards are generally awarded based on actual breakthroughs, whereas the political ones are, let's say, fuzzier.

Dylan16807 4 days ago [-]
> didn't know the Peace Prize was also quite separate from the science based awards.

It's a bit more complicated than that. The committees for Physics and Chemistry (and Economics) are colocated at the Royal Swedish Academy of Sciences. Medicine is elsewhere, as is Literature. Physics, Chemistry, Medicine, and Literature work together for final approval. Peace is completely on its own.

nnm 5 days ago [-]
From my chat with friends who work in the area of drug design, AlphaFold is accurate for overall structure, but is not that accurate for predicting structure around interaction locations.
amrrs 5 days ago [-]
Isn't that quite a big claim? My question, Is that work with Alphafold that significant that it warrants the eyes of Nobel committee? Genuinely curious.
UniverseHacker 5 days ago [-]
In my opinion, absolutely. The "protein folding problem" has been widely regarded as one of the biggest challenges in molecular biology for over half a century, and Alphafold has effectively solved it. I would put this up there with Sanger winning the prize for discovering how to sequence DNA and Kary Mullis for inventing PCR... this will have widespread implications for allowing us to understand, and even design proteins.
dekhn 5 days ago [-]
But they didn't solve the protein folding problem. They solved a simpler problem, protein structure prediction.

What is important about their discovery is that we now know for certain that a judicious combination of expensive-to-obtain structure information, and easy-to-obtain protein sequence relationships can be used to build a generalized protein structure predictor (it can predict structures with no prior example of a fold, although there are limits)... and you don't have solve the general folding problem to do it. You do not need to know the path, to get to the destination!

Many of us in the field expected this to be true but there wasn't any really good example to point to that was widely accepted by the community. And in the ~year or so since this was demonstrated, the community has already found a wide range of uses for this that have validated the structure predictions and demonstrated their utility- using open codes and models.

dhdc 5 days ago [-]
The joke is that the Nobel prize in chemistry is often awarded to non-chemists.
ISL 5 days ago [-]
It could also draw the attention of the Physics Nobel committee. Oodles of physicists have been working on the folding inverse-problem for decades.
epvgwwqe 5 days ago [-]
Seems pretty doubtful. Is there any high impact scientific discovery that AlphaFold directly enabled at this point?
dekhn 5 days ago [-]
The Nobel Prize does not only award scientists for enabling high impact discoveries, but occasionally to people who make a major discovery that has no immediate impact. There is literature dribbling out from folks using AlphaFold models, but that's not what they would be awarding here. This was a long-standing problem that was convincingly solved.
epvgwwqe 5 days ago [-]
If there is precedent for that, then sure they could win.
gone35 5 days ago [-]
Not with recent results in Nature I believe reporting glaring mispredictions. Lots of promotion notwithstanding, AlphaFold may not be usable yet.
gone35 5 days ago [-]
Not Nature but Molecular Systems Biology [2], apologies, in the context of reverse docking for antibiotics discovery:

"We confirm extensive promiscuity, but find that the average area under the receiver operating characteristic curve (auROC) is 0.48, indicating weak model performance."

Derek Lowe had a post about this earlier this month [1] (which includes important qualifications I failed to omit).

[1] "Benchmarking AlphaFold-enabled molecular docking predictions for antibiotic discovery", Molecular Systems Biology 2022 18:e11081 https://doi.org/10.15252/msb.202211081

[2] "Not AlphaFold's Fault", September 7 2022, https://www.science.org/content/blog-post/not-alphafold-s-fa...

flobosg 5 days ago [-]
This is already known. Structural models are usually not well suited for protein-ligand docking, where proper pocket modelling and accurate side chain positioning are key for the determination of true positive hits.
dekhn 4 days ago [-]
Sorry, you don't know what you're talking about. Nobody is truly claiming these models can be used directly for drug discovery (well, some people claim that; but they're wrong). We already knew that, though- the same problem exists with high quality crystal structures.

What would be more interesting is if we did a whole bunch of crystal or cryo-EM structures far from what we've previously determined and demonstrate whether alphafold could make out-of-class predictions for them.

dekhn 5 days ago [-]
Do you have links to that? have not seen those.
gone35 5 days ago [-]
Yes. Posted here (and with some important qualifications): https://news.ycombinator.com/item?id=32948345
j7ake 5 days ago [-]
Congrats to DeepMind! But why did they give the prize to just two members of the alphafold team rather than everyone on the team?

By contrast, the breakthrough prize in physics was awarded to the entire event horizon telescope collaboration for their image of the supermassive black hole.

I would have assumed that the prize for alphafold would also been awarded to the whole team.

KKKKkkkk1 5 days ago [-]
One of the winners, Demis Hassabis, is the head of DeepMind, an org of about 1,000 employees and an annual budget of about $1 bn. One really has to wonder how closely he was involved in the work given his administrative responsibilities.
crakenzak 5 days ago [-]
Completely unrelated, but whenever I see stats like "Deepmind, an org of about 1,000 employees" and then contrast that to the amount of interesting advancements they're making, I cant help but wonder where we'd be with 100,000+ people working on these type of issues.

Kind of how most medical advancements come out of the US, which makes up only <4.25% of world population.

I know orgs dont scale like that and there are hits to productivity with orgs getting less lean but still.

survirtual 4 days ago [-]
Get rid of hierarchies and there might be 100,000 people working on these problems. The masses are limited by the ceiling of greedy egomaniacs. So tired of seeing singular human names take credit for the works of thousands.
xkgt 5 days ago [-]
Is there a good summary of how AlphaFold has made a difference in the few years since it has been released. I was quite excited to hear their success in 2020, so I am curious to see where have we got to in 2 years and what is in the horizon.
GeorgeJIrwin 5 days ago [-]
When will we see the result of this breakthrough in our daily life?

The article mentions:

> So far, the data have been harnessed to tackle problems ranging from antibiotic resistance to crop resilience.

Is any of them is about to be used in our daily life and solve a major problem?

xiphias2 5 days ago [-]
Denis Hassabis has talked about the next evolution of AlphaFold to be developed (and what the team is working on): predicting interactions between proteins. If they are successful (which I really hope they will be as a person with both chronic illness and relatives and friends with illnesses), I can’t think of any drug research where it won’t accerelate the drug development.
TaupeRanger 5 days ago [-]
Except the article is wrong. "Tackle" is doing a lot of work...it doesn't actually mean anything in this case, as AlphaFold has not been shown to help in antibiotic resistance (compounds found using it haven't even been tested in the real world), and has not increased crop resilience in any independent peer reviewed studies or in the real world. It's all still hype at this point.
aardvarkr 5 days ago [-]
“It’s all still hype at this point” implies this is vaporware when it’s a real thing that has solved one of the biggest roadblocks in microbiology. Your claim is analogous to lithium batteries, invented in 1976, taking 25+ years before completely DOMINATING the modern battery market. Science takes time to go from the research stage to mass market adoption. Level set your expectations.
mudrockbestgirl 5 days ago [-]
Survivorship bias. There are 100x more tools and ideas that never got adopted, you just haven't heard of them.
m00x 5 days ago [-]
But AlphaFold is already being used in a ton of research labs at great effect.
TaupeRanger 5 days ago [-]
It is being "used", yes, but to "great effect"? We don't see any "effects" yet - that's the whole point - we don't know yet whether the hype is justified. Time will tell.
m00x 5 days ago [-]
My wife used it in her neuroscience research to see which protein would go through certain tissue barriers (or that's what I understood). It doesn't have the ability to predict interaction with cells and other proteins, and it's not 100%, but it's still useful.

How do you know it's not being useful in research already? Biology is a slow field, you won't see papers mentioning it for months if not years. If I know one person team that uses it, there must be plenty.

TaupeRanger 4 days ago [-]
Sure...we don't know things that haven't happened yet. When a biology paper is published that leads to actual advantages for real life patients or for new technologies to be built that actually solve real world problems, then the hype will begin to be justified. Being "used in research" is not justification for hype. Adding lines to researchers' CVs doesn't actually improve the human condition in any way.
m00x 4 days ago [-]
Well it has been useful for scientists so far, and it's a super early technology. The hype helps researchers know about it, and it will undoubtedly be in papers soon since it's being used today.
mudrockbestgirl 5 days ago [-]
Most of this is probably just marketing/partnerships. It's way too early to tell. The only way to tell whether Alphafold is truly a breakthrough is to wait and see if it passes the test of time.
TaupeRanger 5 days ago [-]
You picked a convenient analogy. No one knew whether lithium batteries were going to be as useful as they were. Much additional testing and work was necessary to prove it. It COULD have failed. Same with Alpha Fold. Abandon your expectations.
krastanov 5 days ago [-]
More resilient crops and reliable antibiotics is "daily life" and "major problem".
bismuthcrystal 5 days ago [-]
Logic compels us to conclude that we will see some results of this on our daily lives and it will become the solution to some "problems".

Problem is, our major problems are mostly social. Biologists will sell you the story that this is a breakthrough that will empower us to improve crop yield and solve world hunger. But we all know we could already have solved it. Turns out the US rather spend billions to build another aircraft carrier instead of develop Africa's farm machinery industry. It is sad. But it is the world.

5 days ago [-]
mellosouls 5 days ago [-]
420official 5 days ago [-]
> ... were recognized for creating the tool that has predicted the 3D structures of almost every known protein on the planet.

I wonder if relying on a tool that doesn't 100% accurately represent reality could have a negative effect on future research

JamesBarney 5 days ago [-]
Current methods are not 100% accurate either. No study is 100%.

Honestly the only field that has a P value that comes close to 100% is physics. Even medicine which is far more rigorous than most fields fails quite often in phase 3 trials after having vetted it in phase 2.

tedsanders 5 days ago [-]
Even physics is nowhere close to “100% accurate.” Most fields of physics approximate many body problems that are infeasible to compute, let alone fully specify. E.g. Astrophysics, solid state physics, nuclear physics, etc. Practitioners regularly use empirically measured parameters like cross-sectional scattering areas, and those parameters are updated and narrowed over time.
dekhn 5 days ago [-]
The predictions made by AlphaFold are now indistinguishable from experimental data collection error so folks aren't super concerned. Anyway structures are typically qualititaive tools useful for thinking about proteins, rather than direct targets of computational predictions (hasn't stopped people from trying).
flobosg 5 days ago [-]
> that doesn't 100% accurately represent reality

It could be argued that this is the case of every scientific tool ever used.

summerlight 5 days ago [-]
If you can develop any kind of non-trivial scientific fields with 100% accuracy and plausible evidence, that finding alone would be qualified for the Nobel prize. Even abstract fields like Math are prone to human errors.
5 days ago [-]
stainablesteel 5 days ago [-]
its a decent enough start because scaling the instrumentation-route of doing this is a lot slower than the ML approach, it can only improve
hanniabu 5 days ago [-]
Does this solve protein folding?
keepquestioning 5 days ago [-]
What's the next AlphaFold? I want to get in early
m00x 5 days ago [-]
BetaFold
refulgentis 5 days ago [-]
Quite funny to me that:

- for the first time, there isn't mountains and mountains of trolling in an Alphafold thread and the comments are _very_ quiet

- the only reason why is a new account tried doing the trolling

- comment is instadead without any manual flagging

- but, people are afraid to post given the one try in 3 hours is dead

Royi 5 days ago [-]
I wonder what people, in 100 years, would say about our era.

One might research, work hard and solve a problem that might change the course of development of a major field and win a recognition by $3M while someone which fills few numbers on a lottery ticket may earn 1-2 folds more.

I wish the system would give this kind of efforts and stories a bigger exposure, recognition and compensation.

Edit: The idea was about the prize amount, not those specific people. It wasn't the best choice, but the idea was that even as a statement, prizes for scientific achievements should be higher so they will be an extreme to all people to recognize and strive for. I guess one could find a better analogy than what I had in mind.

modeless 5 days ago [-]
The people who worked on AlphaFold were (and are) compensated very well. Maybe they didn't win the lottery, but they probably make between 10 and 40 times the median income. And they have received a lot of recognition and exposure, I'd say probably the right amount for the achievement. I'm sure there are issues of this type in the world, but in this case I don't really see a problem.
Drakim 5 days ago [-]
Your point is good but the direction of your contempt is misplaced. Lottery winners accounts for a tiny fraction of people who have unearned and undeserved wealth, and in terms of how many people they screwed over to get to riches, they are like angels in comparison to other rich people.
Royi 5 days ago [-]
I agree with you. I should have made a better choice than lottery. My intention was that the needle which sets the reward for research, long life work pursuing the solution of a problem should move to the right and get those people more.

$3M isn't enough in our days to recognize remarkable work in my opinion. Yes, one of them made a lot of money, but is it true for all the past winners of this prize?

nightski 5 days ago [-]
I'm curious, what would be the right amount in your opinion? How would you value it?
Royi 5 days ago [-]
That would be interesting data driven research.

I can say something about the effect I want to achieve with the money: recognition and publicity. So the amount of money should make them on front page of many news papers and such. Now from data it might interesting to get the right amount.

I just want the people who make such achievements to be recognized by younger generation.

nightski 4 days ago [-]
Not everyone wants those things though. In fact, I could see it being detrimental because then you'd attract a lot of people who only care about those things instead of science itself.

I think we should celebrate scientific achievements as a community. Because the failures are as important as the successes. We shouldn't idolize individuals imho. That is toxic.

random314 5 days ago [-]
Alice walton comes to mind

https://www.mic.com/articles/79039/the-untold-story-of-alice...

And let's not talk about the Sacklers

jonas21 5 days ago [-]
Demis Hassabis made tens, if not hundreds of millions of dollars in the Deep Mind acquisition. I'm sure most people would consider that to be adequate compensation.

If anything, the lesson is that if you care about making lots of money from your research (not everybody does), start a company. And it's easier for academics to start companies today than in any other era.

Royi 5 days ago [-]
I agree, my wording was not perfect.

My point was that such a prize should be backed with more money. Even for the sake of a statement what we consider to be important.

So the emphasize was about the enormous ratio between the two and not about lottery being wrong (Moreover it pays for itself).

5 days ago [-]
derac 5 days ago [-]
The lottery isn't comparable, first of all it's a money raising scheme. I'm sure the alphafold team is well compensated. Almost certainly making high 6 figures. Alphafold got a massive amount of well-deserved coverage as well.
nend 5 days ago [-]
The great wall of china was partially funded by lotteries. I don't think anyone from the future is going to have anything to say about today's lotteries. Lotteries will probably still be popular in a hundred years.
whimsicalism 5 days ago [-]
People are generally compensated for providing goods/services people get value out of. I doubt in 100 years this will be an alien concept.