Google DeepMind’s AlphaFold 3 is available to access by researchers around the world via open-source.
Google DeepMind, the tech giant’s AI research lab, has just open-sourced its AlphaFold 3 protein prediction model. That means academic researchers all over the world can access both the code and training weights for the very first time, following a limited release in May.
What is AlphaFold 3?
Awarded a Novel Price, AlphaFold 3 can predict interactions between proteins and other molecules, such as DNA, RNA, and potential drug compounds. This can open up a wealth of new opportunities, having already mapped more than 200 million protein structures in an unprecedented level of research into structural biology.
That potential is far more accessible now that any researchers can access the code and models, meaning there are vastly more options for how the technology can be put to work.
Before now, other companies like Baidu and ByteDance had already created their versions based on what was published in original research studies and papers. However, the fact that the tech is now accessible to all, regardless of location, financial investment, and so on.
It’s worth noting that DeepMind’s spinoff lab, Isomorphic Labs, maintains exclusive commercial rights, having recently secured $3 billion in pharmaceutical partnerships. However, that doesn’t stop academics putting the code through its paces for research purposes.
This could open plenty more doors in scientific and medical research. Tracking interactions at this level is key for drug discovery and development, with the ability to track behavior across enzymes vital to human metabolism and antibodies that combat infectious diseases.
Making scientific and medical research more accessible is one of the loftier goals for AI, one that OpenAI’s Sam Altman cited in his hopeful letter about the future of artificial superintelligence. Offering wider access to AlphaFold 3 is expected to accelerate groundbreaking discoveries across both biology and medicine, not just by offering access to advanced tech but also by creating an even playing field among researchers outside of powerful institutions who might not normally be able to use such code.
Featured image: Google
