Even After Winning a Nobel Prize, When Will AI Develop Cancer Drugs? [Tech Talk]
Predicting drug candidates alone is not enough
Complex biomolecular simulators are needed
Vast capital is being mobilized to build an "AI laboratory"
In 2024, Demis Hassabis and others who founded DeepMind, Google's artificial intelligence (AI) research lab, received the Nobel Prize in Chemistry. They were recognized for developing AlphaFold, an AI model that predicts three-dimensional protein structures. Protein structures are central to developing new drugs, and predicting new protein structures is a challenge that can take decades or more to solve. AlphaFold prompted optimism that AI could usher in an era in which cancer drugs are developed at tremendous speed.
Yet two years later, AI drug development has still yielded little. How much progress has been made in efforts to combine AI and life sciences? And what obstacles have they encountered?
Nobel Laureate Hassabis Raises $6 Trillion in Funding
Demis Hassabis, founder of Google DeepMind and Isomorphic Labs. Reuters-Yonhap News
View original imageAfter developing AlphaFold, Hassabis founded a new startup, Isomorphic Labs, built around the technology. Like Google DeepMind, which he led, Isomorphic Labs is headquartered in London, UK, and remains highly active. Its mission is to advance AlphaFold and usher in an era in which new drugs can be developed using AI alone.
Isomorphic Labs recently reached a major milestone. On Nov. 17 (local time), it raised a massive $5 billion (approximately KRW 6.67 trillion) from global venture capital (VC) partners, bringing its valuation to more than $50 billion (approximately KRW 66 trillion). The amount ranks among the largest investments ever made in an AI-related company.
However, the drugs Isomorphic Labs is developing with AI have yet to enter even the early stages of development. According to the company's official website, it has obtained "positive efficacy data" in the preclinical stage through collaborations with major pharmaceutical companies.
Why AI Alone Could Not Develop New Drugs
AlphaFold, which demonstrated the potential of AI-based drug discovery. AlphaFold Database website.
View original imageIn fact, "developing new drugs with AI" involves more than simply building a smart AI model. The molecular structures of drug compounds, including proteins, are highly variable in the real world. In other words, AlphaFold could only predict a single still image of a substance that changes constantly. This was a major challenge for companies developing drugs with AI.
That is why pharmaceutical companies developing actual drugs do more than identify drug candidates. They have extensive laboratories and research teams, and spend years continuously formulating and testing drugs. Only then can a drug advance to clinical trials. For AI to test the compounds it predicts, it needs a virtual laboratory inside a computer, much like a pharmaceutical company's wet lab (a specialized facility designed for experiments involving chemicals, proteins, animal cells, and other materials).
Isomorphic Labs Seeks to Build a "Virtual Laboratory" for AI Scientists
Isomorphic Labs, led by Hassabis, no longer relies on AlphaFold alone. The company has already moved on to a major development project called the Isomorphic Labs Drug Design Engine (IsoDDE). IsoDDE is a collection of AI simulators. It functions as a kind of "virtual laboratory for AI," where researchers can test all the real-world variables that affect protein structures, including heat and the minute forces acting between atoms.
Such simulations have traditionally required vast amounts of data and computing power, and have taken a long time to run on expensive supercomputers. Isomorphic Labs, however, aims to replace conventional, resource-intensive simulators with fast, AI-powered ones. While AlphaFold reduced the time needed to predict protein structures, IsoDDE aims to shorten the time required to test drug candidates.
Isomorphic Labs aims to combine numerous simulators into one to create a "virtual laboratory" for AI scientists. Screenshot of the Isomorphic Labs website.
View original imageThe $5 billion in funding is therefore an extremely valuable resource for Isomorphic Labs. Developing AlphaFold was a small experiment back when the work was being done at DeepMind; now, the company must coordinate numerous AI models and computing facilities. In a statement, Isomorphic Labs said, "IsoDDE will push simple predictions to the level of accuracy achieved in real-world experiments," adding, "AI agents built with our technology will navigate highly complex biological systems and find a path to developing new drugs."
Isomorphic Labs will also have its own "real-world infrastructure," including laboratories and research teams for clinical trials. These facilities will help bring drug compounds that exist only on computers into the real world. Isomorphic Labs describes this business model as "a scalable, AI-first translational and clinical research laboratory."
"Gradually, Then All at Once"
The future in which AI produces cancer drugs in an instant is unlikely to arrive anytime soon. IsoDDE will still have to overcome numerous scientific and engineering obstacles before it becomes reliable enough.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Hassabis emphasizes tackling the challenges "one problem at a time." In a post published after the funding announcement, he said, "Drug design using AI required us to rethink the fundamentals," adding, "The time and cost required to develop new drugs will both be dramatically reduced. Gradually, then all at once."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.