edit I was wrong. Please ignore.
> This is a completely new model that was entered in CASP14 and published in Nature.
Alphafold
11–20 of 170 posts
Re: Alphafold
#12edit I was wrong. Please ignore.
From the repo:
> This package provides an implementation of the inference pipeline of AlphaFold v2.0
Re: Alphafold
#13Earlier quoted context omitted.
Nevermind > The simplest way to run AlphaFold is using the provided Docker script. This was tested on Google Cloud with a machine using the nvidia-gpu-cloud-image with 12 vCPUs, 85 GB of RAM, a 100 GB boot disk, the databases on an additional 3 TB disk, and an A100 GPU.
That’s… way closer to consumer than I expected
For inference...
Still accessible, but expensive to run at scale. And training even worse.
Re: Alphafold
#14Ok, so biochemists: which bit of the secret sauce are they leaving out?
Re: Alphafold
#15Fantastic, they released the dataset and code to train the model. Science will be able to proceed.
edit: not the code to train the model, just the code to run inference.
The underlying sequence datasets include PDB strucrures and sequences, and how those map to large collections of sequences with no known structure (no surprise). Each of those datasets represents decades of thousands of scientists work, along with programmers and admins who kept the databases running for decades with very little grant money (funding long-term databases is something NIH hated to do until recently).
Re: Alphafold
#16Alphafold 2 is very very cool, but we need a little dose of reality. It's still a bit away from really solving protein folding as it was marketed.
For example, multi-complex proteins are not well predicted yet and these are really important in many biological processes and drug design:
https://occamstypewriter.org/scurry/2020/12/02/no-deepmind-h...
A disturbing thing is that the architecture is much less novel than I originally thought it would be, so this shows perhaps one of the major difficulties was having the resources to try different things on a massive set of multiple alignments. This is something an industrial lab like DeepMind excels at. Whereas universities tend to suck at anything that requires a directed effort of more than a handful of people.
Re: Alphafold
#17Does anyone on HN work in bio or drug discovery?
Could you give an overview of how people can leverage this (or how you might?).
From reading around about it, it sounds like there's often a need to find a certain type of molecule to activate/inhibit another based on shape and the ability to programmatically solve for this makes the searching way easier.
Is this too oversimplified/wrong? How will this be used in practice.
[Edit]: Thanks for the answers!
Re: Alphafold
#18Fantastic, they released the dataset and code to train the model. Science will be able to proceed. edit: not the code to train the model, just the code to run inference. The underlying sequence datasets include PDB strucrures and sequences, and how those map to large collections of sequences with no known structure (no surprise). Each of those datasets represents decades of thousands of scientists work, along with pr…
There's a preview paper as well: https://www.nature.com/articles/s41586-021-03819-2
Re: Alphafold
#19Fantastic, they released the dataset and code to train the model. Science will be able to proceed. edit: not the code to train the model, just the code to run inference. The underlying sequence datasets include PDB strucrures and sequences, and how those map to large collections of sequences with no known structure (no surprise). Each of those datasets represents decades of thousands of scientists work, along with pr…
> The total download size is around 428 GB and the total size when unzipped is 2.2 TB. Please make sure you have a large enough hard drive space, bandwidth and time to download.
> This was tested on Google Cloud with a machine using the nvidia-gpu-cloud-image with 12 vCPUs, 85 GB of RAM, a 100 GB boot disk, the databases on an additional 3 TB disk, and an A100 GPU.
This is amazingly detailed for a researcher who wants to follow in the track and also Apache licensed, which is one road-bump out of the way for a commercial enterprise, like an actual drug manufacturer who wants to burn some money trying this out.
edit: said the last part too fast, the code has a "the AlphaFold parameters are made available for non-commercial use only under the terms of the CC BY-NC 4.0 license"
Re: Alphafold
#20Fantastic, they released the dataset and code to train the model. Science will be able to proceed. edit: not the code to train the model, just the code to run inference. The underlying sequence datasets include PDB strucrures and sequences, and how those map to large collections of sequences with no known structure (no surprise). Each of those datasets represents decades of thousands of scientists work, along with pr…
There's a preview paper as well: https://www.nature.com/articles/s41586-021-03819-2
Yes, I skimmed the paper already and it wasn't too surprising. There are details that will take some time to parse out to understand how important they are.
Personally, I've found over decades that academic papers like that are far less useful to me than a github project and downloadable data that I can inspect, run and modify on my own. Other folks I know could read that paper and write the code in a day, I always wish I could do that.