TL;DR:
- CZI unveils ambitious mission to “cure, prevent, and manage all disease by the end of the century.”
- CZI announces a monumental computing system with 1,000+ GPUs for AI and LLMs in biomedicine.
- Machine learning tools, like generative AI and LLMs, are reshaping science, technology, and medicine.
- CZI’s CELLxGENE platform aggregates over 50 million single-cell records for researchers.
- AI software tools, including CellGuide, provide vital information on distinct cell types.
- Cryo-electron tomography experiments open doors for understanding protein interactions within cells.
- CZI bridges the gap in computational resources to empower researchers in tackling complex challenges.
- CZI’s focus on enabling tools and technologies accelerates scientific discovery without direct involvement in therapeutics.
Main AI News:
In the realm of scientific innovation, the Chan Zuckerberg Initiative (CZI) is on a mission that echoes the bold ambitions of history. Patricia Brennan, CZI’s Vice President for Science Technology, envisions a future where we can “cure, prevent, and manage all disease by the end of the century.” This visionary goal harkens back to President John F. Kennedy’s iconic challenge in 1961 to land a man on the moon within a decade.
CZI’s commitment to this mission is evident not only in its rhetoric but also in its actions. On September 19, 2023, CZI unveiled plans for one of the world’s largest nonprofit life science research computing systems, boasting over 1,000 GPUs. This supercluster promises to unlock the full potential of artificial intelligence (AI) and large language models (LLMs) in the field of biomedicine.
While CZI’s announcement may not have captivated the world like Neil Armstrong’s moonwalk did in 1969, the potential outcomes are no less groundbreaking.
The Power of the AI Triad
Machine learning (ML) tools, including generative AI and LLMs, have emerged as transformative forces across various domains, from science and technology to pop culture. In the realm of biology and medicine, these tools are already facilitating crucial tasks such as deciphering complex patterns and identifying vital variables within vast datasets. These advancements are propelling drug discovery, genetics, and precision medicine into uncharted territories.
Patricia Brennan succinctly captures the essence of this revolution, noting, “Some say biology is, in many respects, a computational challenge.” Her words resonate with the perspective of the late biophysicist Harold Morowitz, who likened computer science’s role in biology to calculus in physics.
Across all AI applications, whether in biomedicine, robotics, or economics, a consistent triad of elements prevails: data, algorithms, and computing power. CZI has made significant strides in algorithms and data in recent years, particularly in fields like imaging and single-cell biology.
Through its CELL by GENE (CELLxGENE) platform, CZI is fostering collaboration with researchers and the scientific community at large. This initiative aims to aggregate, standardize, integrate, curate, and continually update single-cell data. This democratizes access to a wealth of over 50 million single-cell records, sparing researchers the arduous task of building their datasets from scratch.
“What we’ve seen is that making the entire corpus available and easy to query has spurred the development of new models and research across different areas,” Brennan explains. Researchers are shifting from mere data-level analysis to broader atlas-level investigations, examining tissue atlases and aggregated datasets.
CZ Science research institutes, such as the Chan Zuckerberg Biohub San Francisco and the CZ Imaging Institute, are contributing valuable resources. They have created the protein localization and interaction atlas OpenCell and the cell atlas Tabula Sapiens, along with plans to generate extensive datasets of cells at molecular resolution.
AI-Powered Tools for the Future
CZI’s science technology team is actively involved in AI software development. One standout tool is CellGuide, a free interactive encyclopedia offering essential information on more than 700 distinct cell types and sub-cell types. Developed in collaboration with ChatGPT, CellGuide provides definitions, relevant datasets, ontology tree visualizations, and computational and canonical marker genes.
In tandem, the CZ Imaging Institute is crafting an open-source, cloud-based portal for querying organized data from cryo-electron tomography (cryoET) experiments. This innovative approach holds the promise of uncovering the mysteries of how proteins fit together within cells.
Nicholas Sofroniew, Director of Product Technology at CZI, notes, “In the past five years, I’ve seen tremendous progress in predicting the properties of individual molecules.” However, understanding how these molecules interact within cells remains a frontier ripe for exploration. CryoET measurements, which capture native proteins in their natural environments, could be the next frontier for AI and ML algorithms.
CZI recognizes the vast chasm in computational resources between individual academics and a handful of tech research labs. By bridging this divide, CZI aims to empower researchers to tackle complex problems they might otherwise shy away from.
“There is a unique place we can position ourselves,” says Sofroniew. This positioning spans the types of problems CZI can address, the innovative solutions it can devise, and the computational muscle it can bring to bear on these challenges—constraints increasingly felt in the world of computational biology.
Empowering the Future of Science
CZI’s journey towards this groundbreaking moment has been paved with a dual focus on grantmaking and technology, particularly in the realm of complex data. As CZI continues to gain insights into imaging and single-cell data, it forges ahead with a steadfast commitment to advancing scientific discovery.
While CZI engages in basic research, it also collaborates with researchers, exploring applications and unraveling the mysteries of disease mechanisms. However, CZI’s unique role in realizing its vision of curing, preventing, and managing diseases by the century’s end does not involve directly discovering or developing therapeutics.
As Nicholas Sofroniew explains, “The CZI is focused on the enabling tools, technologies, and data that will advance and empower additional scientific discovery.” Rather than joining the ranks of pharmaceutical companies, CZI’s goal is to accelerate science by building a technological infrastructure that supports data, models, and applications—ultimately propelling scientific progress to new heights.
In a world increasingly reliant on data, algorithms, and computational power, CZI is poised to unify these elements to drive biology’s future. As Patricia Brennan aptly states, “With all the developments, whether it’s in large language models or computational and compute power in recent months or years, we see that there’s an opportunity to bring this all together.” The future of biomedicine is undoubtedly set to reach new horizons with CZI leading the way.
Conclusion:
CZI’s visionary approach to harnessing AI in biomedicine not only empowers researchers but also propels scientific progress. By building an extensive computing infrastructure and democratizing access to critical data, CZI is positioned to reshape the biomedicine landscape, bringing us closer to a future where diseases are conquered. This initiative signifies a significant step toward a more data-driven and technologically advanced era in healthcare and life sciences.
