Would You like a feature Interview?
All Interviews are 100% FREE of Charge
Much of the recent AI hype has centered around captivating digital content generated from simple prompts, while also showing how AI can massively reduce labor and make malicious propaganda more persuasive. There are also concerns about our ability to be what we are. (Fun!) But some of AI’s most promising, and potentially less creepy, work is in medicine. A new update to Google’s AlphaFold software could lead to breakthroughs in research and treatment of new diseases.
The AlphaFold software, developed by Google DeepMind and Isomorphic Labs (also owned by Alphabet), has already demonstrated that it can predict how proteins will fold with shocking accuracy. It catalogs an astounding 200 million known proteins, and Google says millions of researchers have used previous versions in fields such as malaria vaccines, cancer treatments, and enzyme design. He has made discoveries.
Knowing a protein’s shape and structure determines how it interacts with the human body, allowing scientists to develop new drugs or improve existing ones. But the new version, AlphaFold 3, can model other important molecules, including DNA. It can also graph interactions between drugs and diseases, potentially opening exciting new doors for researchers. And Google says it does so with 50% higher accuracy than existing models.
“AlphaFold 3 takes us beyond proteins to a wide range of biomolecules,” says Google’s DeepMind research team I wrote it in a blog post. “This leap could unlock more innovative science, from developing biorenewable materials and more resilient crops to accelerating drug design and genomics research.”
“How do proteins respond to DNA damage? How do they find it and repair it?” John Jumper, Google DeepMind Project Leader Said wired. “We can start answering these questions.”
Before the advent of AI, scientists could only study protein structures through sophisticated methods such as electron microscopy and X-ray crystallography. Machine learning streamlines much of that process by using patterns recognized in training (often imperceptible to humans or standard equipment) to predict the shape of proteins based on their amino acids.
Google says some of AlphaFold 3’s advances come from applying diffusion models to molecular predictions. Diffusion models are a core part of AI image generation tools such as Midjourney, Google’s Gemini, and OpenAI’s DALL-E 3. By incorporating these algorithms into AlphaFold, “the molecular structures that the software generates become clearer.” wired explain. In other words, it takes a configuration that seems vague or ambiguous and makes educated guesses based on patterns in the training data to clarify it.
“This is a huge step forward for us,” Google DeepMind CEO Demis Hassabis said. wired. “This is exactly what you need for drug discovery. You need to see how small molecules bind to drugs, how strongly they bind, and what else they bind to.”
AlphaFold 3 uses a color-coded scale to label the confidence level of predictions, allowing researchers to take appropriate precautions against results that are unlikely to be accurate. Blue means high reliability. Red means less certainty.
Google is developing AlphaFold 3 Free for researchers to use For non-commercial research. However, unlike past versions, the company has not open sourced the project. Professor David Baker of the University of Washington, one of the prominent researchers creating similar software, expressed his disappointment: wired Google has chosen that path. But he was also surprised by the software’s capabilities. “AlphaFold 3’s structural prediction performance is very impressive,” he said.
As for future developments, Google says, “Isomorphic Labs is already working with pharmaceutical companies to apply it to real-world drug design challenges and ultimately develop new treatments that change the lives of patients.” It has said.