Google DeepMind’s latest medical breakthrough borrows a trick from AI image generators

Much of the recent AI hype train has centered around fascinating digital content along with concerns about ability arise from simple instructions destroy the workforce and spread harmful propaganda is more convincing. (Fun!) However, here are some of AI’s most promising and potentially less-failed endeavors sleeps in medicine. New update Google’s AlphaFold software may lead to new disease research and treatment breakthroughs.

AlphaFold software, from Google DeepMind and (also owned by Alphabet) Isomorphic laboratories, has already demonstrated that it can predict how proteins will fold with shocking accuracy. this catalog of an astonishing 200 million known proteinsand Google says millions of researchers have used earlier versions to make discoveries in areas such as malaria vaccines, cancer treatments and enzyme design.

Knowing the shape and structure of a protein determines how it interacts with the human body, allowing scientists to create new drugs or improve existing ones. But a new version, AlphaFold 3, can model other important molecules, including DNA. It can also show interactions between drugs and diseases, which could open interesting new doors for researchers. And Google says it does it with 50 percent better accuracy than existing models.

“AlphaFold 3 takes us beyond proteins to a wide range of biomolecules,” Google’s DeepMind research team wrote in a blog post. “This breakthrough could unlock more transformative science, from developing biorenewable materials and more sustainable plants to accelerating drug design and genomics research.”

“How Proteins Respond to DNA Damage; how do they find it, repair it?” John Jumper, head of the Google DeepMind project he said Wired. “We can begin to answer these questions.”

Before artificial intelligence, scientists could only study protein structures electron microscopes develops such methods X-ray crystallography. Machine learning facilitates much of this process by using patterns recognized from its training (often imperceptible to humans and our standard tools) to predict protein shapes based on amino acids.

Google says that part of AlphaFold 3’s progress comes from applying diffusion models to its molecular predictions. They are central parts of AI image generators such as diffusion models The middle of the journey, Google’s Gemini and OpenAIs DALL-E 3. Incorporating these algorithms into AlphaFold “sharpens the molecular structures generated by the program.” Wired explains. In other words, it takes a shape that appears fuzzy or ambiguous and makes highly educated guesses based on examples from the training data to refine it.

“This is a big breakthrough for us,” said Demis Hassabis, CEO of Google DeepMind Wired. “That’s what you need for drug discovery: you have to see how a small molecule binds to a drug, how potent it is, and what else it can bind to.”

AlphaFold 3 uses a color-coded scale to indicate the level of confidence in its prediction, allowing researchers to exercise appropriate caution with results that are unlikely to be accurate. Blue means high confidence; red means less confident.

Google makes AlphaFold 3 free for researchers for non-commercial research. However, unlike past versions, the company is not open-sourcing the project. David Baker, a professor at the University of Washington, one of the prominent researchers who developed similar software, said he was disappointed Wired Google chose this route. However, he was also impressed with the program’s capabilities. “AlphaFold 3’s structure prediction performance is very impressive,” he said.

As for what’s next, Google says, “Isomorph Labs is already partnering with pharmaceutical companies to apply it to real-world drug design challenges and ultimately develop new treatments that change patients’ lives.”

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *