Omar pfp
Omar
@dromar.eth
Reiterating the Arc Institute is the Bell Labs of our time. Their new STATE model allows in silico prediction of cell perturbations thus visualizing response of drugs at the crllular level. They have also launched a virtual cell challenge to benchmark virtual cell models: https://www.genengnews.com/topics/artificial-intelligence/arc-institute-launches-virtual-cell-challenge-to-accelerate-ai-model-development/   https://arcinstitute.org/manuscripts/State
2 replies
1 recast
15 reactions

HH pfp
HH
@hamud
@ericolsz thoughts?
1 reply
0 recast
0 reaction

Eric Olszewski pfp
Eric Olszewski
@ericolsz
The reason this is funded is because people are just trying to get around the FDA's bullshit. The body is massively complex and we do not know nearly enough about it to model it in silico. They talk about it beating baselines - baselines are absolute dogshit. They are either blind optimists or scammers.
2 replies
0 recast
1 reaction

Omar pfp
Omar
@dromar.eth
I think the goal here is more cellular target, especially assessing gene expression and RNA transcriptomics which is at the foundation of most disease (including cancer as you well know). We still have a poor understanding of many such mechanisms within thr cell which thus nakes it bard to replicate a virtual cell and its response. This is a big leap towards that both in testing it with State and also the data sets Arc released a couple months back. Regarding the drug development process and FDA, I’m not sure how trying to find solutions in shortening and making it for efficient and scalable is useless. Arc isn’t academia, it’s private and funded mostly be tech. A great eloquent read on cell biology and where we are at with it is Song of the Cell by Sid Mukherjee.
1 reply
0 recast
2 reactions

Eric Olszewski pfp
Eric Olszewski
@ericolsz
Okay so I actually misread this - as a means of loading up front of the pipe for drug discovery, that's completely fine and sensible. Sorry I see so much AI BS and thought this was touting a means of doing lead optimization with AI. That said, more leads != more drugs. A lot of things die on the vine and / or don't translate with the BS animal models everyone is pushed to use. I could drone on.
1 reply
0 recast
1 reaction