Alphabet‘s (NASDAQ: GOOG) artificial intelligence (AI) subsidiary Google Deepmind introduced an AI model capable of accurately predicting the structure and interactions of all life’s molecules.
Originally published in the science journal Nature, Google Deepmind announced the breakthrough on Thursday. The discovery promises to illuminate the biological machinery underlying living organisms and to drive breakthroughs in fields ranging from antibiotics and cancer therapy to new materials and resilient crops.
Earlier versions of the DeepMind AlphaFold discovery predicted the 3D structures of 200 million proteins, the building blocks of life, from their chemical constituents. Understanding the shape a protein takes is crucial because it dictates how the protein will function—or malfunction—inside a living organism.
So far, millions of researchers worldwide have utilized AlphaFold 2 to make discoveries in areas such as malaria vaccines, cancer treatments, and enzyme design. Scholars have cited AlphaFold more than 20,000 times, and many prizes, including the most recent Breakthrough Prize in Life Sciences, have recognized its scientific impact.
AlphaFold 3 advances beyond proteins to encompass a broad spectrum of biomolecules. This leap could enable more transformative science in developing biorenewable materials, enhancing crop resilience, accelerating drug design, and advancing genomics research.
Given an input list of molecules, AlphaFold 3 generates their joint 3D structure, revealing how they fit together. It models large biomolecules such as proteins, DNA, and RNA, as well as small molecules, also known as ligands — a category that includes many drugs.
Additionally, AlphaFold 3 can model chemical modifications to these molecules, which regulate healthy cellular functions and can lead to disease when disrupted.
Read more: Microsoft creates generative AI bot tailored for American intel agencies
Read more: Liberty Defense Holdings provides security for major Canadian airport
AlphaFold uses diffusion network similar to AI image generators
The capabilities of AlphaFold 3 stem from its next-generation architecture and training that now encompasses all life’s molecules. At the heart of the model lies an improved version of the Evoformer module — a deep learning architecture that propelled AlphaFold 2’s remarkable performance. After processing the inputs, AlphaFold 3 uses a diffusion network to assemble its predictions, similar to those in AI image generators. This diffusion process starts with a cloud of atoms and, over many steps, converges to its final, most accurate molecular structure.
“AlphaFold continues to get better, and increasingly more relevant for biological investigations,” said Paul Nurse, the Nobel Prize-winning geneticist and chief executive and director of the London-based biomedical research center the Francis Crick Institute.
“This third version will enable increased accuracy in predicting the structures of complexes between different macromolecules, as well as associations between macromolecules, small molecules and ions.”
AlphaFold 3’s predictions of molecular interactions surpass the accuracy of all existing systems. As a single model that computes entire molecular complexes in a holistic manner, it uniquely unifies scientific insights.
In their Nature article, the researchers describe how AlphaFold3 predicts the interactions of proteins with other proteins, ions, strands of genetic code, and smaller molecules, such as those developed for medicines. In tests, the program’s accuracy ranged from 62 per cent to 76 per cent.
Google DeepMind also launched AlphaFold Server, which helps scientists around the world perform non-commercial research. Biologists can use the abilities of AlphaFold 3 to model structures composed of DNA, RNA, proteins as well as a series of ions, chemical modifications and ligands.
Read more: Verses AI onboards chief product officer in push for AI product Genius
Read more: Verses AI raises CAD$10M in private placement and leans into AI product, Genius
Artificial intelligence has proven transformative
Artificial intelligence has blown up since its introduction back in 2023 and has since moved on past OpenAI’s large language models into something truly transformative.
According to a report from Grand View Research, the global artificial intelligence market size was valued at USD 62.35 billion in 2020 and is expected to expand at a compound annual growth rate (CAGR) of 40.2 per cent from 2021 to 2028.
Beyond Google DeepMind, plenty of companies have gotten onto the AI bandwagon in the time since OpenAI introduced Chat GPT, including some of the biggest names. For example, Microsoft (NASDAQ: MSFT) is a supporter of OpenAI. Also, Amazon (NASDAQ: AMZN) teamed with Anthropic on its Claude AI competitor to ChatGPT.
Then there’s smaller companies looking to make inroads with the technology.
A few of these include Nextech3D.AI (CSE: NTAR) which recently launched an AI powered 3D model search engine, Liberty Defense Holdings (TSXV: SCAN) (OTCQB: LDDFF) (FRANKFURT: LD2 A ), which incorporates AI to improve the security of public spaces like airport security and professional sports venues. Also, there’s VERSES AI Inc. (CBOE: VERS) (OTCQB: VRSSF)
VERSES built its Genius API with a slightly different approach from other companies using AI. Instead of choosing an artificial path to intelligence, the company proposes a more natural path. In this case, it’s one grounded in physics with a deep understanding of biological intelligence—which are first principles guided by the wisdom of genius of nature.
Verses AI is a sponsor of Mugglehead news coverage
.
Follow Joseph Morton on Twitter
joseph@mugglehead.com
