AlphaFold 3: Google DeepMind Seeks to Transform Healthcare
Google DeepMind this week (May 2024) announced AlphaFold 3, its new AI model that can predict the structure and interactions of all life’s molecules.
Hoping to operate with precision accuracy, the system will witness how molecules - including DNA and proteins - interact together across millions of different combinations. The goal, according to Google DeepMind, is to truly understand the processes of life up close.
“We hope AlphaFold 3 will transform our understanding of the biological world and drug discovery,” the company highlighted in its press release. “This leap could unlock more transformative science, from developing biorenewable materials and more resilient crops, to accelerating drug design and genomics research.”
AlphaFold 3 builds on previous versions of the AlphaFold system that was originally launched in 2020.
Science up close: Getting AI under the microscope
Once provided with an input list of molecules, AlphaFold 3 works by generating their 3D structure to reveal how they fit together. The new system is able to model large biomolecules, including proteins, DNA and RNA and can also model chemical modifications to them.
The AI has been trained on all of life’s molecules, which makes it even more advanced. After processing the inputs it has been given, AlphaFold 3 can assemble predictions using a diffusion network - much like those in AI image generators.
Google DeepMind suggests that AlphaFold 3 is able to predict molecular interactions at a far more improved accuracy than all other existing systems. It is able to unify scientific insights, no matter the complexity, and is already demonstrating a 50% improvement compared to existing prediction methods. As a result, Google DeepMind has successfully doubled prediction accuracy.
Scientists will be able to access the majority of AlphaFold 3’s capabilities for free via the company’s newly launched research tool, AlphaFold Server.
Likewise, Isomorphic Labs - founded by Google DeepMind co-founder and CEO Sir Demis Hassabis - is collaborating with pharmaceutical companies to apply the AI to real-world drug design challenges, with the aim of developing life-changing treatment options for patients.
AlphaFold continues to power medical advancements
The initial AlphaFold AI system was launched as a protein structure database. It predicts the 3D structure of a protein from its amino acid sequence. Harnessing the power of AI, researchers can now analyse the data obtained from the protein in a much shorter space of time.
The system was hailed as revolutionary upon its release, offering greater opportunities for innovation and hopefully leading to treatments for diseases and improved patient outcomes around the world.
To date, millions of researchers globally have used AlphaFold 2, the most recent iteration of AlphaFold, to make discoveries in areas including malaria vaccines, cancer treatments and enzyme design. The AI system has been cited more than 20,000 times.
As a whole, the healthcare industry has been no stranger to AI in recent months. Companies like Google DeepMind have been developing the technology for early cancer detection, more proactive care and freeing up healthcare professionals to handle more complex tasks.
If harnessed responsibly, the application of AI in healthcare could be transformative for the industry.
******
Make sure you check out the latest edition of AI Magazine and also sign up to our global conference series - Tech & AI LIVE 2024
******
AI Magazine is a BizClik brand
