Google DeepMind and Isomorphic Labs have unveiled their new artificial intelligence (AI) model AlphaFold 3, which they say could “transform our understanding of the biological world and drug discovery”.
AlphaFold 3 is detailed in a paper published in the prestigious journal Nature.
The new model is said to have a substantially updated diffusion-based network compared to its predecessor AlphaFold 2.
The authors say that AlphaFold 2 has already spurred “a revolution” in modelling protein structures – a vital and challenging aspect of developing new drugs.
It has been used by millions of researchers in areas as diverse as designing new enzymes, developing new treatments for cancer and working on new malaria vaccines.
It has been cited more than 20,000 times in scientific papers, with breakthroughs enabled through the technology winning numerous prizes, including the Breakthrough Prize in Life Sciences.
AlphaFold 3 is claimed to be able to predict the structure and interactions of “all life’s molecules” with far more accuracy than any existing method, including AlphaFold 2.
The AlphaFold team said that it offers a 50% accuracy improvement compared to existing prediction techniques, using a benchmark known as PoseBusters.
New AI architecture allows a broader range of biochemical predictions
They added that the new model goes beyond proteins to work with a broader spectrum of molecular systems, significantly increasing its versatility.
These molecular systems include multiple proteins, DNA, RNA, small molecule ligands and ions.
Potential use cases could include more effective new drug design, research in genomics and even biorenewable materials.
Many of AlphaFold 3’s capabilities will be made available to researchers for free through a new platform known as AlphaFold Server, as long as they are working on non-commercial projects.
The team said that the platform is relatively easy to use, allowing biologists and other researchers to utilise the AI without the need for expertise in machine learning.
The platform also allows them to access the technology with limited hardware and computational resources.
The team acknowledged that there are always potential risks with such a powerful tool but said that they had engaged with numerous experts and third parties across biosecurity, research and industry, in order to better understand and mitigate such risks.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
