For the purposes of this question, “released” means that sufficient information is publicly available to essentially reproduce results in the white paper published by Isomorphic Labs on October 31, 2023 and to extend this to equivalent predictions: https://www.isomorphiclabs.com/articles/a-glimpse-of-the-next-generation-of-alphafold
This will resolve “YES” even if free-of-charge use is limited to non-commercial use, requires some form of user registration, and/or requires using an online service (no local execution) with a modest limit on prediction rate.
1,000🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ116 | |
| 2 | Ṁ107 | |
| 3 | Ṁ106 | |
| 4 | Ṁ95 | |
| 5 | Ṁ42 |
People are also trading
Damn, questions condition perfectly predicted it :p : it is free for non commercial use, with google account registration, and using an online server
https://golgi.sandbox.google.com/about
On second read, this is a bit ambiguous because this definitely does not include arbitrary small molecule ligand docking
@CamillePerrin Yes so far it will resolve "NO" based on "reproducing the white paper" based on limited selection of ligands.
Today's announcement is more or less what I'd predicted when writing the question and I was on the fence about extent of ligand/PTM selection.
As far as the question predicting (perhaps; not over yet) the outcome -- reading between the lines of simultaneously promoting open research aspects (which are great and incredibly valuable -- I use AFDB content almost every day) and over-inflating biosecurity concerns to justify digging the moat. There's a pretty narrow range of likely outcomes at that point. Announcing the pharma deal made it more certain, but I guessed the point of the white paper was a sales pitch that would have that outcome in the first place.
White paper:
Ligands on server:
A non-commercial server with limited capability (such as restrictions on ligands) does not count as "released" in my opinion. Hundreds of papers have used the downloaded AlphaFold2 and AlphaFold-Multimer code to develop new pipelines and benchmark them and apply them in a high throughput manner to protein-protein interactions, antibodies, kinases and GPCRs in different states, etc. None of that will be available for AlphaFold3 is DeepMind/Isomorphic Labs don't release the code to non-commercial users.
@RolandDunbrack While I agree on the principle, for the purposes of this question I went with something I considered about 50/50 likely at the time I wrote the question. That was before I'd heard about the Eli Lilly and Novartis deals, so now I figure that if I participated in my own market I would buy below 50% and I would buy at about 50% if the date was extended to the end of 2024.
FYI, although it’s hard to see how this can be subjective, I will not participate in the market. Also FWIW I’m not sure which way I would bet if I did. I hope this resolves YES since the product is immediately useful for useful science.
Edit: I was also away from the news and posed this question before reading about Isomorphic Labs’ deals with Eli Lilly and Novartis.