Cryo-electron tomography (cryoET) is a powerful imaging technique that captures high-resolution, 3D views of intact cellular structures. But while 3D tomograms can be created in a matter of days, interpreting them often requires labor-intensive manual annotation that can drag on for months, resulting in a daunting backlog of cryoET data. Now, though, researchers at the CZ Imaging Institute have combined machine learning tools and large-scale computing to dramatically accelerate the process, annotating more than 13,000 tomograms in just 3.5 days. Read more.
Researchers at CZ Biohub SF have developed a software tool that allows scientists to tackle complex bioimage processing without the need to write code. Omega is a Large Language Model-based conversational agent implemented as a plug-in for napari, a popular open-source image viewer. Loïc Royer, Sr. Group Leader & Director of Imaging AI at CZ Biohub SF, described Omega in a Nature Methods paper last month. The plug-in, which has been available to researchers for about a year, is already being downloaded around 2,000 times per month. Read more.
How lung microbiomes can be predictive of mortality
Bone marrow transplants are sometimes the only option for children battling leukemia and related conditions. But though the procedure can be life-saving, it leaves the patient’s immune system temporarily out of commission and can result in dangerous infections that may cause fatal lung injury. Now, a new Nature Medicine study co-led by Matt Zinter of UCSF and Biohub SF President Joe DeRisi demonstrates how metagenomic analysis of the lung microbiome can predict which children are most vulnerable during this risky but potentially life-saving procedure. Read more.
Kavli Prize goes to CZ Biohub Network board member
Paul Alivisatos, president of the University of Chicago and a CZ Biohub Network board member was one of three awardees of the 2024 Kavli Prize in Nanoscience, an international award honoring “outstanding researchers doing fundamental science that moves the world forward.” Alivisatos was selected for his work on quantum dots, tiny semiconductor crystals whose colors vary depending on their size. In the 1990s, Alivisatos and his research group at UC Berkeley became the first to apply quantum dots to bioimaging, an approach now used by researchers around the world in methods like labeling, live cell tracking, and in vivo imaging. Read more.
life / science
STORIES FROM THE CZ BIOHUB NETWORK
Meet the Director of Computational Biology at Biohub SF
Yasin Şenbabaoğlu was just a high school student in Istanbul when he discovered his passion for merging mathematics and biology. That drive led to a master’s degree in statistics followed by a Ph.D. in bioinformatics, and, eventually, to a role at Genentech, where he developed AI tools to help predict cancer patients’ responses to immunotherapy. After many years using quantitative sciences to tackle health questions, he’s now the Director of Computational Biology at Biohub SF.
UChicago News: Scientists invent “living bioelectronics” that can sense and heal skin
Bioelectronics — like pacemakers, cochlear implants, and brain–machine interfaces — integrate biological systems and electronic circuits to improve health. But getting electronics to work well with biological tissue remains a major challenge, a problem University of Chicago Professor Bozhi Tian has built his lab around. Now, in a study supported by a CZ Biohub Chicago Acceleration Program Award and published in Science, Tian’s group has created a “living bioelectronics” prototype. The device incorporates microbes that naturally occur on human skin, and was successful in reducing psoriasis-like symptoms in mice. Read more.
Columbia Engineering News: Gordana Vunjak-Novakovic to advance cancer research with Investigator award
Columbia University engineer Gordana Vunjak-Novakovic has been named one of seven inaugural Investigators at CZ Biohub NY. Well known “for her pathbreaking work in engineering functional human tissue for use in regenerative medicine,” Vunjak-Novakovic plans to investigate reciprocal interactions between bone marrow and tumor cells, in part by creating “a human tissue–based ‘organ-on-chip,’’' Holly Evarts writes for Columbia Engineering’s news site. Read more.
Immunity: AI and immunology
The journal Immunity asked nine experts to share their perspectives about AI’s current and future impact on immunology research. Check out what Biohub NY Advisory Committee member John Tsang had to say about what’s working — like predicting protein structure and molecular interactions — and what still needs to be done to keep moving forward. Read more.
Recent publications
Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity
Science, Houlahan et al.
Biohub SF Investigator Cristina Curtis believes some tumors are born to be bad. Here, Curtis and other Stanford researchers demonstrate that our germline genomes — the gene sequences we inherit from our parents — play a major role in the potential for cells with possible cancer-causing mutations to evade our immune defenses. “The study challenges the dogma that most cancers arise as the result of random mutations that accumulate during our lifetimes,” writes the Stanford Medicine News Center in a summary of the findings. Read more.
Network-based elucidation of colon cancer drug resistance mechanisms by phosphoproteomic time-series analysis
Nature Communications, Rosenberger et al.
Dysregulation of cellular signaling cascades plays a critical role in cancer, and the ability of cancer cells to “rewire” signaling networks in response to drugs creates a pathway for treatment resistance. Here, CZ Biohub New York President Andrea Califano and colleagues introduce Virtual Enrichment-based Signaling Protein-activity Analysis (VESPA), “an algorithm designed to elucidate mechanisms of cell response and adaptation to drug perturbations,” and use it to investigate drug resistance mechanisms in colorectal cancer cells. Read more.
Robust virtual staining of landmark organelles
bioRxiv, Liu et al.
Researchers at CZ Biohub SF are advancing methods for virtual cell staining — an approach that eliminates the cost and delays of traditional staining while also avoiding the physical alteration of the cells under study. Here, they report model training protocols that lead to robust virtual staining of nuclei and cell membranes in multiple cell types. The work was a collaborative effort from the teams of Platform Leader Shalin Mehta and Group Leaders Adrian Jacobo, Carolina Arias, and Manuel Leonetti. Read more.
Self-Healing Materials for Bioelectronic Devices
Advanced Materials, Liu et al.
CZ Biohub Chicago scientists Claire Liu, Daniel Wang, and Shana Kelley review the history, design principles, and future outlook of self-healing materials, essential in wearable bioelectronic devices that enable various forms of health monitoring. Read more.
A pathologist–AI collaboration framework for enhancing diagnostic accuracies and efficiencies
Nature Biomedical Engineering, Huang et al.
A team led by CZ Biohub SF Investigator James Zou and Zhi Huang, both of Stanford University has developed a “human-in-the-loop” framework called nuclei.io to overcome limitations of off-the-shelf AI models for pathology. They found that when examining histology images for colorectal cancer and endometriosis, nuclei.io was better and faster than humans or AI working alone. Read more.
Permittivity tensor imaging: modular label-free imaging of 3D dry mass and 3D orientation at high resolution
Nature Methods, Li-Hao Yeh et al.
A biomolecule’s “permittivity tensor” describes how the phase and polarization of light change as it passes through the molecule. Along with collaborators, CZ Biohub SF researchers led by Platform Leader Shalin Mehta have developed a new technique to use light microscopy to capture this information and create 3D images of biomolecules without the need for labels or dyes. The team found that permittivity tensor imaging (PTI) outperforms previous label-free imaging techniques for similar measures, despite using simpler hardware. The publication comes with open-source resources to help others adopt PTI. Read more.
We’re hiring! Check out our open positions in Chicago, New York, and the San Francisco Bay Area.
The Chan Zuckerberg Biohub Network is a group of nonprofit research institutes that bring together physicians, scientists, and engineers with the goal of pursuing grand scientific challenges on a 10- to 15-year time horizon. These institutes partner with Chan Zuckerberg Science in its goal to understand the mysteries of the cell and how cells interact within systems. This collaboration will bring us closer to our mission to cure, prevent, or manage all disease by the end of the century. To learn more, visit www.czbiohub.org.
Chan Zuckerberg Biohub, 499 Illinois St., San Francisco, CA 94158
