Astera Institute

Public isn’t the same as usable. We’re proud to be a founding partner of Project ProtoStar, a new initiative through @huggingface, helping make federally funded datasets, models and code AI-ready so more researchers can build on them.
huggingface.co/project-protos…

The @scientistsorg Metascience Fellowship puts scientists inside federal agencies to experiment with how science gets funded and what counts as impact.

If you know someone who should do this work, send them the link to apply. fas.org/career/metasci…

Today, we are announcing our third cohort of residents! Each resident will receive up to $2M to work on a high impact scientific problem, with the governing principle being that they build something people actually use and build it openly.

Read more about Radial's support of @PrismScienceOrg from @beckypferdehirt. asterainstitute.substack.com/p/radial-doubl…

The last decade brought a dramatic convergence of lab experiments on protein structures and AI models designed to make sense of them. But in the wake of the cryo-EM boom and AlphaFold, structural biology is in a tricky place.

How do we keep learning from static protein structures when real proteins are actually in constant motion?