2 months ago
26
Artificial intelligence still has much to achieve. However, last year, it received significant recognition for advancing biological sciences, thanks to Google DeepMind and its AlphaFold protein structure prediction technology, which played a crucial role in winning the 2024 Nobel Prize in Chemistry.
But that’s merely the initial step in a long, ambitious journey to leverage AI to advance life sciences. In a fireside chat at the World Economic Forum (WEF) in Davos 2025, Demis Hassabis, CEO of Google DeepMind, stated, “Ultimately, my dream would be to simulate a virtual cell.”
In another interview, he stated that AlphaFold provided a static interpretation of a protein, emphasising that one can only ‘truly’ understand what is occurring if the dynamics and interactions between different components within a cell are considered.
“A virtual cell project is about building an AI simulation of a full working cell,” he said, he noted, suggesting he might start with a simple organism like a yeast cell.
Virtual Cell May Arrive Five Years From Now
At the World Economic Forum, Hassabis was accompanied by Ardem Patapoutian, a distinguished molecular biologist who was awarded the Nobel Prize in Physics or Medicine in the year 2021. Patapoutian elaborated on the idea of a virtual cell, saying, “Seeing the cell as is and where things are, both with respect to the architecture and 3D space, is very, very interesting.”
He also mentioned that this aids us in achieving a deeper understanding of human biology beyond elevated levels of protein expression.
“Let’s say a certain protein is expressed very highly in a cell. But when you actually look at it and see it, all of it is localised at the tip of the neuron where something very specific is happening,” he added.
Hassabis also stated that this could be crucial for understanding how a cell reacts to the injection of a specific nutrient or drug. At present, this process is carried out ‘painstakingly’ in a laboratory, but according to him, this project would enable researchers to perform it a million times faster and at a lower cost.
However, he noted that a laboratory would still be required as a final step to validate the predictions and conduct clinical trials to assess the efficacy of the research. Hassabis mentioned that the virtual cell project may materialise within five years from now.
That said, this is one of many instances where researchers have investigated the concept of a virtual cell.
How to Build an AI Virtual Cell
In December of last year, Stanford University, along with several prominent global universities, released a research study examining how artificial intelligence can create a virtual cell.
The researchers outline three capabilities that an AI virtual cell should possess. It should be able to create a universal representation across species and cell types, help accurately predict cell functions, allow computer-based experiments to test hypotheses, and guide data collection to expand its capabilities.
The proposed model will consist of two components: an AI model to interpret the biological systems and ‘Virtual Instruments’ to analyse and simulate the results.
The researchers said that this has the potential to help cancer researchers study how mutations cause healthy cells to become malignant, developer biologists to explore how cell lineages change with disruptions, and microbiologists to examine the effects of viral infections on both the cells and their host organisms.
However, while one question asks ‘how to build a virtual cell’, the authors consider the more important one ‘how to trust in their competence’.
“To this end, a comprehensive and adaptable benchmarking framework will be needed,” said the authors.
“It [AI virtual cell] must account for dynamic distributions that evolve due to environmental changes, infections, genetic variants, and other such factors causing distribution shifts,” they added.
Moreover, the authors mentioned another challenge: addressing bias in the data used to train the model and ensuring its diversity.
“As virtual cell efforts mature, the dialogue between the scientists who develop models, those who generate experimental data, and funding organisations must be further intensified,” said the authors.
However, they also emphasise the importance of using open data resources and standardised data formats to ensure that the project contributes to the overall benefit of humanity. This would also require a healthy collaboration between investors, bioethics experts, regulators, academia, researchers, and every other stakeholder involved.
All things considered, all of these ambitions project towards increasing the longevity of life.
Long Live..Longevity Tech?
“We don’t have an unambiguous metric for progress. If you must pick one, we can make a strong case for life expectancy,” reads a line from the book ‘The Anthology of Balaji’, a collection of statements made by Balaji Srinivasan, an entrepreneur who has built multiple genomics companies.
Srinivasan opines that if the purpose of technology is to reduce scarcity, then the ultimate purpose of technology is to eliminate mortality, which is the main source of scarcity.
Several technologists have embraced longevity technology. The recently announced $500 billion Project Stargate aims to eliminate diseases and improve longevity.
According to Larry Ellison, CTO of Oracle and one of the companies funding the project, the infrastructure to be built aims to enable AI to “create cancer vaccines, personalised medicine, and pandemic prevention.”
“We will see diseases get cured at an unprecedented rate. We will be amazed at how quickly we’re curing cancer and heart disease…” said Sam Altman, CEO of OpenAI, the company that is at the forefront of The Project Stargate.
The company also has ambitions to build longevity tech, and Altman ‘personally funded’ $180 million to Retro Biosciences, a longevity science company, in 2023. Recently, Retro Biosciences also announced a $1 billion fundraising round to back clinical trials for three drugs.
Most tech leaders share the same opinion. Dario Amodei, CEO of Anthropic, said, “If you think about what we might expect humans to accomplish in an area like biology in 100 years, I think a doubling of the human lifespan is not at all crazy,” while speaking at the WEF, in Davos.
That said, we’re undoubtedly far away from realising such ambitions. In an interaction with AIM, Paras Chopra, founder of Turing’s Dream, who is also a gold medal recipient for his bachelor’s in biotechnology, believes that people are blowing up the ambition of immortality.
“I’m all up for living longer and fitter, but I think we are up against very hard physical and biological limits when we’re talking about living forever,” said Chopra.
“As technologists, we have a tendency to get super excited and super optimistic about things,” he added.
