Morgridge Institute for Research

Data intensive research continues to drive innovation and discovery across Earth system science (ESS). ESS datasets maintained in science discipline specific repositories, including climate model projections, historical reanalysis products, and observational datasets, provide rich resources to support these initiatives. While significant progress has been made through the hosting of datasets by commercial cloud providers, many of these data resources– sometimes stored in non-standard formats–are primarily maintained in unconnected, domain-focused data systems designed to support the legacy “download, clean and analyze model”. This is a time consuming process with bandwidth and storage requirements that may be prohibitive, particularly for institutions with limited resources. This combination of the download, clean, and analyze model, and the use of non-standard formats, combine to create a barrier to realizing the full research potential of ESS data assets. This presentation will highlight the National Science Foundation National Center for Atmospheric Research (NSF NCAR) efforts to develop and deploy its Geoscience Data Exchange, Research Data Commons (GDEX, https://gdex.ucar.edu). GDEX is designed to overcome the challenges described above by: 1) curating standards based (FAIR), Analysis and AI optimized (AR/AI) versions of global and regional atmospheric reanalysis outputs, earth systems simulation outputs, and observations produced at NSF NCAR and partner organizations, 2) providing direct access to these datasets through its integration with on-premise computational resources, and 3) providing performant distributed access through its integration the Open Science Data Federation’s (OSDF, https://osg-htc.org/services/osdf). The OSDF supports streaming data access and integration with a variety of data and compute services through its system of geographically distributed data caches, including commercial cloud hosted open datasets. GDEX’s integration with OSDF supports a wider variety of cross-domain research use cases by enabling efficient access to the spectrum of datasets hosted through OSDF’s origin access points. Finally, GDEX is integrated with colocated data analytics services to support rapid development and iteration of data science (e.g. AL/ML) workflows, and facilitate open sharing of those workflows. To promote user adoption of these services, an example set of reference data analysis workflows have been seeded in public collaboration software repositories and documented in JupyterBook style web pages. GDEX users are encouraged to submit their own workflow examples through this resource, amplifying the impact of their science by allowing others to more easily build upon their work.

Abstract Human eosinophils activated in suspension with IL5 or IL33 undergo morphological change prior to adhesion. Refractive granules, which contain major basic protein-1 and other toxic proteins, move to one side of the cell, the granulomere, and the two nuclear lobes move to the other. How these features persist when eosinophils become adherent and migrate is not known. We now compare behavior of activated eosinophils on surfaces coated with ITGAM/ITGB2-integrin ligands fibrinogen or periostin using live cell imaging of reporters of tubulin/actin organization and cell viability. We find that unlike eosinophils activated with IL5, IL33-activated eosinophils undergo two stages of activation; a preliminary pear-like activation in which the cell develops polarity, followed by a flattening of the eosinophil into a thin pancake-like morphology with less discrete polarization. IL5-treated eosinophils migrated persistently for more than an hour with nucleopod in the back. In contrast, IL33-treated eosinophils moved more slowly and within 30 min transitioned to a flattened morphology with nuclear lobes in the center and dispersed motile granules. Loss of cell viability after an hour, although variable, in all comparisons was greater among IL33-treated eosinophils on periostin. We sought to understand how cytoskeletal elements may drive these differences in morphology. Cytoskeletal elements had similar responses when activated with IL5/IL33; vimentin collapsed from a web-like network at the periphery of the cell and condensed adjacent to the nucleopod/nuclear interface, f-actin was found in the granulomere as well as the tip of the nucleopod and forward periphery, and microtubules radiated from the microtubule organizing center (MTOC) spanning both the nucleopod and the granulomere. The dynamic formation of microtubules correlated with cellular locomotion, suggesting mesenchymal migration within these cells. These in vitro findings suggest that adhesion plays an important role in determining functional morphology and demonstrates new insights into IL33-activated eosinophils. This work suggests roles for activators and adhesive substrates in regulating the behavior of activated eosinophils in tissues.

Variation is a foundational biological principle that has historically been marginalized—both due to limited experimental accessibility and because of idealized, stereotypic blueprints rooted in essentialist thinking. With the advent of genetics and quantitative biology investigating environmental influences on the phenotype, variation was redefined from mere noise to a fundamental property. Modern light sheet microscopy now enables high-resolution, long-term imaging of dynamic processes across large populations, making it possible to systematically study phenotypic variation in vivo. Yet, the resulting high-dimensional datasets overwhelm traditional modes of analysis and visualization, risking the loss of biological insight. The transition from qualitative representation to quantitative measurement demands new epistemic practices—shifting from selective human interpretation to computational abstraction. Instead of relying on either very limited sampling or exhaustive scanning, we advocate for representative sampling of phenotypic variation: adaptive, model-guided systems that dynamically sample biological variation using real-time feedback, directing attention towards biologically relevant events and rare or extreme phenotypes. The underlying models act as the interface to human insight, constructing navigable, queryable representations of variation as a multidimensional manifold shaped by genetics, environment, and stochasticity. Crucially, adaptive systems call for new methods of visualizations—interfaces that encode uncertainty, consensus, and distributional structure. Such visualizations should preserve the interpretability of historical illustrations while fully embracing biological variation. The future of biology lies not in acquiring more data, but in developing smarter ways to sample, represent, and understand it.

One-size-fits-all opioid treatment for chronic pain may undermine treatment effectiveness and cause adverse effects. We explore personalized nonpharmacological and noninvasive chronic pain treatment as a promising approach to minimize these risks and optimize outcomes for each patient. Our paper introduces the definition of personalized treatment, emphasizes the significance of intervention personalization, and outlines strategies for inter- and intra-patient personalization in chronic pain management. We also discuss empirical studies on personalized chronic pain treatment. Research has shown a variety of strategies for personalizing nonpharmacological and noninvasive chronic pain treatment. These strategies have demonstrated the potential to improve effectiveness. However, many critical scientific questions remain unanswered. To fully harness the potential of personalized nonpharmacological and noninvasive chronic pain treatment, future research should conduct robust clinical trials and meta-analyses comparing personalized and non-personalized chronic pain interventions, examine diverse personalization approaches, incorporate ecological momentary assessment into outcomes assessment, leverage digital technologies, engage patients in treatment design, and assess factors influencing the implementation of personalized chronic pain treatment.

Accurately quantifying cellular morphology at scale could substantially empower existing single-cell approaches. However, measuring cell morphology remains an active field of research, which has inspired multiple computer vision algorithms over the years. Here, we show that DINOv2, a vision-transformer based, self-supervised algorithm, has a remarkable ability for learning rich representations of cellular morphology without manual annotations or any other type of supervision. We apply DINOv2 to cell phenotyping problems, and compare the performance of resulting models, called Cell-DINO models, on a wide variety of tasks across two publicly available imaging datasets of diverse specifications and biological focus. Compared to supervised and other self-supervised baselines, Cell-DINO models demonstrate improved performance, especially in low annotation regimes. For instance, to classify protein localization using only 1% of annotations on a challenging single-cell dataset, Cell-DINO performs 70% better than a supervised strategy, and 24% better than another self-supervised alternative. The results show that Cell-DINO can support the study of unknown biological variation, including single-cell heterogeneity and relationships between experimental conditions, making it an excellent tool for image-based biological discovery.

Exponential increases in microbial and viral genomic data demand transformational advances in scalable, generalizable frameworks for their interpretation. Standard homology-based functional analyses are hindered by the rapid divergence of microbial and especially viral genomes and proteins that significantly decreases the volume of usable data. Here, we present Protein Set Transformer (PST), a protein-based genome language model that models genomes as sets of proteins without considering sparsely available functional labels. Trained on >100k viruses, PST outperforms other homology- and language model-based approaches for relating viral genomes based on shared protein content. Further, PST demonstrates protein structural and functional awareness by clustering capsid-fold-containing proteins with known capsid proteins and uniquely clustering late gene proteins within related viruses. Our data establish PST as a valuable method for diverse viral genomics, ecology, and evolutionary applications. We posit that the PST framework can be a foundation model for microbial genomics when trained on suitable data.

Extracellular matrix (ECM) plays fundamental roles in animal development, regeneration, and disease. The difficulty of tagging endogenous matrix proteins in vertebrates has limited the understanding of ECM composition and dynamics in complex tissues. To visualize vertebrate ECM components, we tagged zebrafish Laminin, gamma 1 (Lamc1), Collagen, type I, alpha 2 (Col1a2), and Transforming growth factor, beta-induced (Tgfbi) using C-terminus in-fusion genome editing. Analysis of these knock-in lines revealed distinct expression of each protein in various tissues during development and regeneration. Fluorescent recover after photobleaching (FRAP) analysis further indicated that Lamc1 is stable in fin fold matrix but more dynamic in myoseptal matrix of developing zebrafish, while Col1a2 and Tgfbi are stable matrix components in myosepta. Strikingly, we found that Col1a2-mScarlet protein accumulates at the amputation plane during tailfin regeneration, where it remains concentrated for several days and distant from the regeneration blastema. This “foundation” region also displayed a distinct transcriptome suggesting active and dedicated events at the base of the regenerating appendage. Our resource enables live capture of ECM dynamics that can identify new events in developing and regenerating zebrafish.

Abstract Description Systemic lupus erythematosus (SLE) is a chronic autoimmune disorder which primarily affects young women. SLE has heterogenous clinical manifestations and is associated with increased mortality and reduced quality of life. Studies have shown that T lymphocytes are implicated in SLE pathogenesis and are the target of hydroxychloroquine (HCQ) and Mycophenolate Mofetil (MMF), the most common medications used to manage SLE. These medications may inhibit T cell activation, a process that is regulated by metabolism. Here, we use autofluorescence lifetime imaging microscopy (FLIM) of metabolic cofactors NAD(P)H and FAD as a non-destructive method to assess metabolic changes in T cells from SLE patients without staining or fixation. Single T cell metabolic heterogeneity was investigated within and between SLE patients (n = 15) of different sexes, ages, disease severities, and treatment regimens. Preliminary findings indicate a significant decrease in the long lifetime components of both NAD(P)H and FAD and a significant increase in the proportion of free NAD(P)H and protein-bound FAD in SLE patients of high and low disease severity compared to healthy donors. Overall, autofluorescence FLIM could provide a rapid single-cell assessment to monitor SLE disease progression and predict flares, thereby guiding treatment. Funding Sources the Morgridge Institute for Research, the Daniel M. Albert Chair, NIH/NCI R01 CA278051 Topic Categories Basic Autoimmunity (BA)

The fields of bioimaging and image analysis are rapidly expanding as new technologies transform biological questions into novel insights. While professionals of varying expertise are essential to achieving these advancements, early-career scientists—a prominent user group within the imaging community—are often assumed to have the prerequisite knowledge and ability to use these tools. This demographic, consisting of students, post-docs, and bioimage analysis trainees, is critical for the field to continue to evolve and flourish. However, obstacles such as geographic location, language barriers, insufficient funding or training, and instrument availability hinder access to resources and introduce significant knowledge gaps, especially for scientists in early-career stages. Democratized resources for bioimaging and analysis such as forums, community organizations, and publicly available datasets have been helpful in overcoming barriers to access for early-career scientists. Here, we discuss the current tools and resources available for early-career researchers, highlight their limitations from the learners’ perspective, and propose strategies to better support this group. As bioimage analysis extends broadly into many scientific disciplines, we implore all members of this community, regardless of experience level, to empower next-generation scientists.

ABSTRACT Melanoma continues to be the deadliest form of skin cancer, emphasizing the need for new therapeutic strategies. Targeting tumor metabolism, particularly oxidative phosphorylation (OXPHOS) has emerged as a promising approach due its role in melanoma tumor survival, metastasis, and treatment resistance. In this study, we investigate the metabolic and antitumor effects of atovaquone, an FDA-approved and safe OXPHOS inhibitor that has not been previously tested in melanoma. We show that atovaquone rapidly and effectively inhibits OXPHOS and impairs glycolysis in melanoma, leading to metabolic reprogramming observed via metabolic imaging and marked depletion in energy (ATP) stores. Atovaquone also induces oxidative stress, evidenced by increased reactive oxygen species levels, DNA damage, and upregulation of antioxidant proteins. Moreover, atovaquone reduced melanoma cell viability and migration in vitro , and slowed tumor growth in vivo . Notably, these effects were observed in both BRAF -wild-type and mutant melanoma models, suggesting its potential as an effective treatment across different subtypes. Our study identifies atovaquone as a metabolic disruptor with antitumor activity against melanoma, supporting further investigation as a repurposed therapeutic strategy.