The Remarkable Discovery of the Norn Cell
In the late 19th century, French doctor François Jolibois made a fascinating observation while studying blood samples from the Andes: the red blood cells crucial for carrying oxygen had increased by a staggering 42%. This discovery opened the door to a deeper understanding of the human body's remarkable ability to produce essential cells on demand when necessary.
It took scientists nearly 70 years to isolate the hormone responsible for this capability, which they named erythropoietin or "red maker" in Greek. The next breakthrough came decades later when researchers in Israel identified a unique kidney cell, dubbed the "Norn cell," responsible for producing this vital hormone in response to oxygen shortages.
AI's Groundbreaking Discovery of the Norn Cell
The journey to uncover the Norn cell was a long and arduous one, spanning over 134 years. However, last summer, computers in California managed to identify this elusive cell in just six weeks, thanks to the power of artificial intelligence (AI).
The Stanford researchers behind this discovery set up computers to self-learn biology, using an AI program similar to ChatGPT, which had learned language from billions of internet texts. Instead of feeding the AI pre-labeled data, the team provided raw information on millions of actual cells, including their chemical and genetic profiles, and let the computers analyze the data independently.
The outcome was remarkable: the AI model was able to identify the Norn cell, a previously unseen cell type, as one of over a thousand distinct cell types. "It's incredible because the AI model discovered the Norn cell in the kidney without being informed of its existence," said Jurovich, the computer scientist behind the project.
The Rise of Foundation Models in Biology
The Norn cell discovery is just one example of the transformative impact of AI-powered "foundation models" in the field of biology. These models, which are trained on vast amounts of data, are uncovering new insights into gene functions, cell development, and the underlying mechanisms of life.
One such model, called Gene Forger, was developed by Dr. Christina Theodoris, a medical resident at Boston Children's Hospital. By training the model on 30 million cells from 106 human studies, Gene Forger gained a profound understanding of gene behavior across various cell types. The model was able to accurately predict the effects of deactivating certain genes in specific heart cell types, a feat that would have been challenging for human researchers to achieve.
Another foundation model, called Universal Cell Embedding (UCE), was developed by the Stanford team. This model was trained on 33 million cells from the Cell Atlas, one of the largest cell databases globally. UCE was able to categorize over a thousand cell types by observing patterns in gene activation, echoing discoveries made by generations of biologists. Furthermore, the model demonstrated the ability to apply its knowledge to unfamiliar species, accurately identifying cell types in species like the naked mole rat.
The Potential and Challenges of Foundation Models
The success of foundation models in biology has sparked both excitement and cautious optimism. These models have the potential to uncover mysteries about cancer, other diseases, and even the fundamental nature of life itself. Dr. Eric Topol, the head of the Scripps Research Translational Institute, believes that "discovering something about biology that biologists haven't been able to would be a landmark moment, and I believe it's coming."
However, these models are not without their limitations. Computational biologist Kazia Kedra and her team at Oxford University have put foundation models like Gene Forger and SCPT through rigorous testing, challenging them with unseen cell atlases. While the AI models excelled at certain tasks, they sometimes underperformed compared to simpler software, highlighting the need for caution in their application.
Dr. Lovich, the Stanford researcher, acknowledges that the volume of data in cell atlases is relatively small compared to the vast information used to train language models like ChatGPT. He expresses a desire for an "internet of cells," where larger and more diverse cell data is available to enhance the foundation models' understanding of cell functionality.
The Future of Biology and the Potential Risks
As foundation models continue to evolve, experts envision a future where these AI-powered tools could revolutionize biology. Bo Wang, the creator of SCPT and a computational biologist at the University of Toronto, believes that a comprehensive mathematical model of a cell could enable scientists to simulate experiments digitally, predicting cell behavior without the need for physical laboratories.
Dr. Quake, a researcher, speculates that foundation models might uncover not only the cell types known to exist but also those that could potentially exist, adhering to the biochemical rules of life. He imagines creating a "map of life's boundaries," exploring beyond which life cannot sustain, and even devising chemical blueprints to convert regular cells into ones with unique capabilities, such as clearing plaque from blood vessels or examining diseased organs from the inside.
However, these advancements also introduce new risks. Over 80 biologists and AI specialists have called for regulations to prevent the potential misuse of this technology, such as the development of novel biological weapons derived from artificially created cells. Concerns about privacy and the exposure of sensitive genetic information also arise as foundation models become tailored to individual genomes.
The success of foundation models has also prompted a re-evaluation of the role of human biologists in research. While traditional biology has valued the creativity and labor involved in experimental research, computers analyzing vast quantities of cell data might reveal complex patterns and insights in a fraction of the time, challenging our notions of scientific creativity and the future role of biologists in research.
Conclusion: The AI Revolution in Biology
The integration of AI technologies in genetic research and cellular biology represents a pivotal moment in medical science, promising revolutionary advances in diagnostics and therapeutic strategies. The remarkable discoveries made by foundation models, such as the identification of the Norn cell, offer fresh perspectives on disease treatment and the intricate workings of human cells, marking a significant leap forward in our understanding of health and biology.
While the potential of these AI-powered tools is vast, the challenges and risks they introduce must be carefully navigated. As the field of biology continues to evolve, the collaboration between human researchers and AI-driven foundation models will be crucial in unlocking the secrets of life and paving the way for groundbreaking advancements in medicine and beyond.
.jpg)
0 Comments