Experiments must be in the loop. Automation helps, but the real bottleneck is choosing the right optimisation targets and defining what “good” looks like.

AlphaFold succeeded because protein structures could be validated experimentally. Not all biological tasks have that luxury. 5/8

The second direction: post-AlphaFold biology. Architectures built specifically for molecules, not NLP models.

To replicate AlphaFold’s impact, we need:
• high-quality experimental benchmarks
• sufficient data volume
• new architectures
• tight integration with wet-lab validation 5/8

The capabilities are real:

• AlphaFold: 214M protein structures (Nobel Prize 2024)
• Insilico Medicine: AI drug to Phase IIa in 30 months vs 4-6 years
• IDx-DR: First FDA-approved autonomous AI diagnostic

But here's what matters more: implementation.

Thanks! Through the NSF NAIRR pilot program we were paired with an industry partner to support our AI-related project. The program connected us with NVIDIA and they provided three months of access to a 256-GPU DGX cluster, which is where we ran all of the AlphaFold predictions.

Outstanding - would be great to better understand NVIDIA’s supporting role.

Prof Ho's fellowship project will combine the capabilities of specialized scientific AI models like AlphaFold with more general AI models that can reason about science without understanding it deeply, building a stepping stone toward scientific artificial general intelligence.

Quantum isn't hype—it's buildable.
Follow @TheTechWorldPod for AI/quantum breakdowns, no-buzz scripts.
#AI #Quantum #WillowChip #AlphaFold spti.fi/dRbvXjf

User annotations are visible on the 2D and 3D tracks, but they are only applied for the duration of the user session and not saved.

The AlphaFold Database provides open access to over 240 million protein structure predictions. The resource is a collaboration between EMBL-EBI and Google DeepMind.

Hi Anni, have you tried our bioinfirmatics AI assistant Pipette.bio? It can access AlphaFold predictions for a given protein, along with several other protein databases.

We next turned to AlphaFold to understand how MTR interacts with its binding partners. AlphaFold of MTR with its rescue partner MTRR threw a surprise our way: AF3 suggested two regions of interaction with one newly suggested that deviated from the previously known interaction with MTRR’s FMN domain.