The J. David Gladstone Institutes

Less than a decade ago, a slew of startups like Metagenomi and Mammoth Biosciences raised huge sums of money to advance CRISPR gene editing systems found in nature. Now, the combination of artificial intelligence models with genetic and biological data has made those original pitches seem far more trivial. In a paper publishing Thursday in Science , more than 20 scientists at the Arc Institute and Stanford University shared details on Evo, an AI biology model that designed a novel CRISPR system from scratch. The Arc Institute, based in Palo Alto, CA, gives guaranteed, unrestricted budgets to affiliated researchers, which cover a wide range of areas from genomics to AI to neuroscience. Evo is a first step in a wildly ambitious vision of simulating, predicting and ultimately controlling biology. The paper also detailed Evo-made genetic modifiers called transposons, and includes the first examples of a language model co-designing these protein-RNA and protein-DNA complexes, the model’s developers said. Evo was trained on 2.7 million genomes of prokaryotes and phages, using a new AI architecture called StripedHyena to simultaneously process and consider more text than other AI systems can handle. It was first made public this February through a preprint and blog post. The Science paper details new experiments like the CRISPR and transposon examples. The model’s gene editing system, dubbed EvoCas9-1, was as good at cutting DNA as the Nobel Prize-winning original Cas9 system used in the first approved CRISPR therapy. The team fed the Evo model with over 72,000 CRISPR-Cas complexes to help it generate new sequences. EvoCas9-1 was ultimately 73% similar in its amino acid sequence compared to the original Cas9 enzyme, according to the paper. Other AI bio groups have pursued similar research, most notably Profluent Bio, which publicized its OpenCRISPR system earlier this year . Both Evo and Profluent’s results are similar in novelty to original CRISPR, but the Arc Institute’s Patrick Hsu and Brian Hie noted Evo was able to reach a new level of generalization. Specifically, Profluent trained three models to make its gene editing contraption: one model focused on the DNA-cutting enzyme, another designed the guide RNA, and a third predicted how compatible those two parts are together. Evo, by contrast, designed the entire system as a single model. While the early results are impressive, Evo is limited to the realm of tiny organisms like bacteria. It’s unclear if it will work in understanding more complex organisms, like humans. And even on the scale of CRISPR, it remains to be seen if it can design gene editing systems superior to ones found in nature, such as by making them smaller, more potent or more precise. “Future work will determine whether the StripedHyena architecture is the optimal choice for large eukaryotic genomes, which can be up to 160 billion nucleotides and may require larger context lengths to learn the grammar of distant regulatory interactions,” Gladstone Institutes researcher Christina Theodoris wrote in a related Perspective piece . Still, Evo-designed tools could be useful additions to the molecular biology toolbox, Hsu and Hie wrote in an email, listing examples like nucleases, reverse transcriptases, and ligases. The Arc team is already working on Evo2, hoping to scale to more complex organisms and take leaps in modeling more than just DNA, RNA and proteins — going to pathways, cells and organs, they added. “Further development of large-scale biological sequence models like Evo, combined with advances in DNA synthesis and genome engineering, will accelerate our ability to engineer life,” the Evo authors wrote in the paper. Editor’s note: This story has been updated to reflect some of the paper’s scientists are affiliated with Stanford University, but not the Arc Institute.

SAN FRANCISCO & MONTREAL--( BUSINESS WIRE )--Nomic Bio Inc. (Nomic), a proteomics company with the highest-throughput platform for multiplexed protein quantification, and the Parker Institute for Cancer Immunotherapy (PICI), a collaborative consortium of the world’s leading immuno-oncology experts, today announced a research collaboration to quantify circulating protein biomarkers that are associated with patient responses to immunotherapy across cancer types. Under the collaboration, Nomic will profile blood samples from PICI’s RADIOHEAD (Resistance Drivers for Immuno-Oncology Patients Interrogated by Harmonized Molecular Datasets) cohort, a prospective longitudinal study of 1,070 patients receiving standard-of-care immune checkpoint inhibitor (ICI) treatment regimens in 52 community hospitals across the United States. The testing, which will be conducted with Nomic’s nELISA platform, will provide parallel quantification of hundreds of proteins related to cancer and immune cell responses. The results will assess the levels of inflammatory and cancer-related proteins and explore their association with patient response to therapy, survival rates, and likelihood of developing immune-related adverse events. “By characterizing the soluble signaling molecules of our longitudinal patient samples, we aim to elucidate the communication pathways among immune cells in response to disease states and treatments,” said Tarak Mody, PhD, chief business officer at PICI. “The data from this mission-focused collaboration, along with other omics-data as part of the RADIOHEAD program, will enhance our understanding of treatment resistance and accelerate advancements in precision medicine to improve patient care.” This project will benefit from Nomic's revolutionary nELISA platform which provides industry-leading specificity and delivers absolute quantification of hundreds of key protein targets. These advancements combined with unparalleled throughput have already established Nomic as a leading proteomics solution provider in drug discovery. “We are excited to work with PICI to help elucidate how cancer patients respond to immunotherapy, as it becomes a first-line treatment option across an increasing number of cancer types,” said Milad Dagher, CEO and Co-Founder of Nomic. “Since its inception, the raison d’être of Nomic has been to empower scientists with the data they need to improve patients’ lives. This collaboration is a great opportunity for us to apply our nELISA technology to a project that can eventually help healthcare providers make more informed decisions about life-saving immunotherapies.” Nomic and PICI intend to begin publishing data generated by the study in early 2025. About Nomic At Nomic, we are making protein profiling broadly accessible. At the heart of our platform is the nELISA™, a proprietary technology that enables massive multiplexing with true flexibility, full quantification, and unmatched cost-efficiency. Nomic has developed the most comprehensive solution for inflammation research that allows for the high-throughput quantification of hundreds of secreted proteins. Nomic is rapidly expanding its biological footprint to cover proteins across the entire proteome, unlocking new possibilities for diverse applications, uses cases, and disease areas. For more information, visit nomic.bio and follow the company on LinkedIn. About the Parker Institute for Cancer Immunotherapy The Parker Institute for Cancer Immunotherapy (PICI) is radically changing how cancer research is done. Founded in 2016 through a $250 million gift from Silicon Valley entrepreneur and philanthropist Sean Parker, with an additional $125 million commitment announced in 2024, the San Francisco-based nonprofit is an unprecedented collaboration between the country’s leading immunotherapy researchers and cancer centers. PICI Network research institutions include Stanford Medicine; the University of California, Los Angeles; the University of California, San Francisco; the University of Pennsylvania; Dana-Farber Cancer Institute; Gladstone Institutes; and Weill Cornell Medicine. PICI also supports top researchers at other institutions, including The University of Texas MD Anderson Cancer Center, Memorial Sloan Kettering Cancer Center, City of Hope, Fred Hutchinson Cancer Research Center, Icahn School of Medicine at Mount Sinai, Institute for Systems Biology and Washington University School of Medicine in St. Louis. By forging alliances with academic, industry and nonprofit partners, PICI makes big bets on bold research to fulfill its mission: to accelerate the development of breakthrough immunotherapies to turn all cancers into curable diseases. Find out more at parkerici.org and follow us on LinkedIn , X (formerly Twitter) @parkerici , and on Spotify and Apple Podcasts. About RADIOHEAD The RADIOHEAD (Resistance Drivers for Immuno-Oncology Patients Interrogated by Harmonized Molecular Datasets) program is a pan-tumor, prospective cohort study of 1000+ immunotherapy naive individuals on standard of care immune checkpoint inhibitor (CPI) treatment regimens. This study captured a comprehensive profile of each participant by collecting blood samples and clinical features at pretreatment, early on-treatment, and late timepoints from 52 community oncology clinics across the United States. If participants experienced immune-related adverse events, additional samples were collected upon presentation and in follow-up visits. In all, over 3700 longitudinal samples were collected, with a focus on major IO indications such as non-small cell lung cancer (~1400 samples) and malignant melanoma (~500 samples). Pairing comprehensive multi-omic profiling of these longitudinal samples with associated demographic and clinical outcomes provides an opportunity to identify clinically actionable mechanisms for CPI resistance and adverse events, discover targets for combination therapies and post-CPI treatment, and inform system biology approaches to elucidate disease pathways.

Plus, news about the Gladstone Institutes, ImmunityBio, Keen Vision Acquisition Corporation and Medera: Basilea Pharmaceutica earns $25M milestone from Pfizer: The payment was triggered after the antifungal Cresemba exceeded a sales threshold as part of a licensing agreement with the drugmaker. The agreement covers Cresemba’s sales in Europe and 16 countries in Asia and China. Cresemba pulled in $489 million in global sales between April 2023 and March 2024, according to Basilea. — Katherine Lewin Connect Biopharma reduces workforce in China: As part of the company’s decision to become more US-centric, Connect cut its China-based workforce by 15% over the past year, and it expects to further reduce its workforce this year. It had 81 full-time employees at the end of last year. It has cash and cash equivalents of $110 million, which gives the company runway into the first half of 2027. — Katherine Lewin Gladstone Institutes announces $350M capital campaign: The organization plans to use the money it raises to add more than 300 scientists and 15 laboratories as part of a goal to “develop the next generation of biotechnologies and advance discoveries to the clinic as rapidly as possible.” The labs will add 100,000 square feet of space. — Katherine Lewin ImmunityBio lays off 20 people in California and Colorado: The cuts were detailed in a WARN report submitted to the Florida Department of Commerce on Aug. 30, with manufacturing and quality control-related teams mostly impacted. The company’s IL-15 agonist Anktiva won FDA approval in April to treat bladder cancer. ImmunityBio did not immediately respond to a request for comment. — Max Bayer Keen Vision Acquisition Corporation, Medera to merge: Medera, a privately-held biotech, focuses on developing gene and cell therapies for difficult-to-treat and currently incurable diseases. Its three most advanced candidates are gene therapy treatments for heart failure indications. Keen is a SPAC that went public in February 2023. The companies expect the merger to close in the fourth quarter. At that time, the company will be known as Medera. Keen will get $149.5 million in cash proceeds from the transaction. — Katherine Lewin