A spatially resolved timeline of the human maternal-fetal interface.

Beginning in the first trimester, fetally derived extravillous trophoblasts (EVTs) invade the uterus and remodel its spiral arteries, transforming them into large, dilated blood vessels. Several mechanisms have been proposed to explain how EVTs coordinate with the maternal decidua to promote a tissue microenvironment conducive to spiral artery remodelling (SAR)1-3. However, it remains a matter of debate regarding which immune and stromal cells participate in these interactions and how this evolves with respect to gestational age. Here we used a multiomics approach, combining the strengths of spatial proteomics and transcriptomics, to construct a spatiotemporal atlas of the human maternal-fetal interface in the first half of pregnancy. We used multiplexed ion beam imaging by time-of-flight and a 37-plex antibody panel to analyse around 500,000 cells and 588 arteries within intact decidua from 66 individuals between 6 and 20 weeks of gestation, integrating this dataset with co-registered transcriptomics profiles. Gestational age substantially influenced the frequency of maternal immune and stromal cells, with tolerogenic subsets expressing CD206, CD163, TIM-3, galectin-9 and IDO-1 becoming increasingly enriched and colocalized at later time points. By contrast, SAR progression preferentially correlated with EVT invasion and was transcriptionally defined by 78 gene ontology pathways exhibiting distinct monotonic and biphasic trends. Last, we developed an integrated model of SAR whereby invasion is accompanied by the upregulation of pro-angiogenic, immunoregulatory EVT programmes that promote interactions with the vascular endothelium while avoiding the activation of maternal immune cells.

DeepCell Kiosk: scaling deep learning-enabled cellular image analysis with Kubernetes.

Deep learning is transforming the analysis of biological images, but applying these models to large datasets remains challenging. Here we describe the DeepCell Kiosk, cloud-native software that dynamically scales deep learning workflows to accommodate large imaging datasets. To demonstrate the scalability and affordability of this software, we identified cell nuclei in 106 1-megapixel images in ~5.5 h for ~US$250, with a cost below US$100 achievable depending on cluster configuration. The DeepCell Kiosk can be downloaded at https://github.com/vanvalenlab/kiosk-console ; a persistent deployment is available at https://deepcell.org/ .

Experimental assessment of elemental analyzer isotope ratio mass spectrometry normalization methodologies for environmental stable isotopes.

RATIONALE: In stable isotope mass spectrometry, isotope delta values are normalized to internationally recognized reference scales using a combination of certified and in-house isotope reference materials. Numerous techniques exist for performing this normalization, but these methodologies need to be experimentally assessed to compare their effect on reproducibility of isotope results. METHODS: We tested normalization methods by the number of reference materials used, their matrix, their isotope range, and whether normalization required extrapolating beyond the isotope range. We analyzed eight commercially available isotope reference materials on a ThermoFinnigan Delta-V isotope ratio mass spectrometer (IRMS) and an Elementar VisION IRMS for nitrogen and carbon isotope composition via solid combustion with an elemental analyzer and computed every possible isotope normalization (n = 612). Additionally, we assessed how sample matrix affected linearity effects on both instruments using five in-house reference materials. RESULTS: Normalizations exhibited the best performance when the reference materials spanning an isotope range of at least 20‰ were matrix matched with the samples and did not require extrapolation beyond the calibration curve. When these conditions were not met, the number of reference materials used had a significant effect on accuracy, with normalizations composed of two reference materials exhibiting particularly inconsistent performance at isotope ranges below 20‰. Linearity effects were found to exceed instrument precision by two orders of magnitude irrespective of matrix type and were not predicted by working gas diagnostics. CONCLUSIONS: Interlaboratory comparability of isotope results is improved when operators of elemental analyzer isotope ratio mass spectrometry (EAIRMS) systems select reference materials spanning an isotope range of at least 20‰. Additionally, using three or more isotopic reference materials, avoiding extrapolation beyond the range of the normalization curve, and matching the matrix of the reference materials to the samples improve normalizations.

ΔSCOPE: A new method to quantify 3D biological structures and identify differences in zebrafish forebrain development.

Research in the life sciences has traditionally relied on the analysis of clear morphological phenotypes, which are often revealed using increasingly powerful microscopy techniques analyzed as maximum intensity projections (MIPs). However, as biology turns towards the analysis of more subtle phenotypes, MIPs and qualitative approaches are failing to adequately describe these phenotypes. To address these limitations and quantitatively analyze the three-dimensional (3D) spatial relationships of biological structures, we developed the computational method and program called ΔSCOPE (Changes in Spatial Cylindrical Coordinate Orientation using PCA Examination). Our approach uses the fluorescent signal distribution within a 3D data set and reorients the fluorescent signal to a relative biological reference structure. This approach enables quantification and statistical analysis of spatial relationships and signal density in 3D multichannel signals that are positioned around a well-defined structure contained in a reference channel. We validated the application of ΔSCOPE by analyzing normal axon and glial cell guidance in the zebrafish forebrain and by quantifying the commissural phenotypes associated with abnormal Slit guidance cue expression in the forebrain. Despite commissural phenotypes which display disruptions to the reference structure, ΔSCOPE was able to detect subtle, previously uncharacterized changes in zebrafish forebrain midline crossing axons and glia. This method has been developed as a user-friendly, open source program. We propose that ΔSCOPE is an innovative approach to advancing the state of image quantification in the field of high resolution microscopy, and that the techniques presented here are of broad applications to the life science field.

A Foundation Model for Cell Segmentation.

Cells are a fundamental unit of biological organization, and identifying them in imaging data - cell segmentation - is a critical task for various cellular imaging experiments. While deep learning methods have led to substantial progress on this problem, most models in use are specialist models that work well for specific domains. Methods that have learned the general notion of "what is a cell" and can identify them across different domains of cellular imaging data have proven elusive. In this work, we present CellSAM, a foundation model for cell segmentation that generalizes across diverse cellular imaging data. CellSAM builds on top of the Segment Anything Model (SAM) by developing a prompt engineering approach for mask generation. We train an object detector, CellFinder, to automatically detect cells and prompt SAM to generate segmentations. We show that this approach allows a single model to achieve human-level performance for segmenting images of mammalian cells (in tissues and cell culture), yeast, and bacteria collected across various imaging modalities. We show that CellSAM has strong zero-shot performance and can be improved with a few examples via few-shot learning. We also show that CellSAM can unify bioimaging analysis workflows such as spatial transcriptomics and cell tracking. A deployed version of CellSAM is available at https://cellsam.deepcell.org/.

Ecological characteristics impact PFAS concentrations in a U.S. North Atlantic food web.

This is the first comprehensive study of per- and polyfluoroalkyl substances (PFAS) in a coastal food web of the U.S. North Atlantic, in which we characterize the presence and concentrations of 24 targeted PFAS across 18 marine species from Narragansett Bay, Rhode Island, and surrounding waters. These species reflect the diversity of a typical North Atlantic Ocean food web with organisms from a variety of taxa, habitat types, and feeding guilds. Many of these organisms have no previously reported information on PFAS tissue concentrations. We found significant relationships of PFAS concentrations with respect to various ecological characteristics including species, body size, habitat, feeding guild, and location of collection. Based upon the 19 PFAS detected in the study (5 were not detected in samples), benthic omnivores (American lobsters = 10.5 ng/g ww, winter skates = 5.77 ng/g ww, Cancer crabs = 4.59 ng/g ww) and pelagic piscivores (striped bass = 8.50 ng/g ww, bluefish = 4.30 ng/g ww) demonstrated the greatest average ∑PFAS concentrations across all species sampled. Further, American lobsters had the highest concentrations detected in individuals (∑PFAS up to 21.1 ng/g ww, which consisted primarily of long-chain PFCAs). The calculation of field-based trophic magnification factors (TMFs) for the top 8 detected PFAS determined that perfluorodecanoic acid (PFDA), perfluorooctane sulfonic acid (PFOS), and perfluorooctane sulfonamide (FOSA) associated with the pelagic habitat biomagnified, whereas perfluorotetradecanoic acid (PFTeDA) associated with the benthic habitat demonstrated trophic dilution in this food web (calculated trophic levels ranged from 1.65 to 4.97). While PFAS exposure to these organisms may have adverse implications for ecological impacts via toxicological effects, many of these species are also key recreational and commercial fisheries resulting in potential for human exposure via dietary consumption.

Tidal Flushing Rather Than Non-Point Source Nitrogen Pollution Drives Nutrient Dynamics in A Putatively Eutrophic Estuary.

The effects of nonpoint source nutrients on estuaries can be difficult to pinpoint, with researchers often using indicator species, monitoring, and models to detect influence and change. Here, we made stable isotope measurements of nitrogen and carbon in sediment, water column particulates, primary producers, and consumers at 35 stations in the reportedly eutrophic Barnegat Bay (New Jersey) to assess N sources and processing pathways. Combined with water quality and hydrological data, our C and N isoscapes revealed four distinct geographic zones with diverging isotopic baselines, indicating variable nutrient sources and processing pathways. Overall, the carbon stable isotopes δ13C) reflected the terrestrial-marine gradient with the most depleted values in the urban and poorly flushed north of the estuary to the most enriched values in the salt marsh-dominated south. In contrast, the nitrogen stable isotope values δ15N) were most enriched near the oceanic inlets and were consistent with offshore δ15N values in particulate organic matter. Several biogeochemical processes likely alter δ15N, but the relatively lower δ15N values associated with the most urbanized area indicate that anthropogenic runoff is not a dominant N source to this area. Our findings stand in contrast to previous studies of similar estuaries, as δ15N signatures of biota in this system are inversely correlated to population density and nutrient concentrations. Further, our analyses of archival plant (Spartina sp., Phragmites australis) and shell (Geukensia demissa, Ilyanassa obsoleta) samples collected between 1880 and 2020 indicated that δ15N values have decreased over time, particularly in the consumers. Overall, we find that water quality issues appear to be most acute in the poorly flushed parts of Barnegat Bay and emphasize the important role that oceanic exchange plays in water quality and associated estuarine food webs in the lagoon.

Whole-cell segmentation of tissue images with human-level performance using large-scale data annotation and deep learning.

A principal challenge in the analysis of tissue imaging data is cell segmentation-the task of identifying the precise boundary of every cell in an image. To address this problem we constructed TissueNet, a dataset for training segmentation models that contains more than 1 million manually labeled cells, an order of magnitude more than all previously published segmentation training datasets. We used TissueNet to train Mesmer, a deep-learning-enabled segmentation algorithm. We demonstrated that Mesmer is more accurate than previous methods, generalizes to the full diversity of tissue types and imaging platforms in TissueNet, and achieves human-level performance. Mesmer enabled the automated extraction of key cellular features, such as subcellular localization of protein signal, which was challenging with previous approaches. We then adapted Mesmer to harness cell lineage information in highly multiplexed datasets and used this enhanced version to quantify cell morphology changes during human gestation. All code, data and models are released as a community resource.

Characterizing the diverse cells that associate with the developing commissures of the zebrafish forebrain.

During embryonic development of bilaterally symmetrical organisms, neurons send axons across the midline at specific points to connect the two halves of the nervous system with a commissure. Little is known about the cells at the midline that facilitate this tightly regulated process. We exploit the conserved process of vertebrate embryonic development in the zebrafish model system to elucidate the identity of cells at the midline that may facilitate postoptic (POC) and anterior commissure (AC) development. We have discovered that three different gfap+ astroglial cell morphologies persist in contact with pathfinding axons throughout commissure formation. Similarly, olig2+ progenitor cells occupy delineated portions of the postoptic and anterior commissures where they act as multipotent, neural progenitors. Moreover, we conclude that both gfap+ and olig2+ progenitor cells give rise to neuronal populations in both the telencephalon and diencephalon; however, these varied cell populations showed significant developmental timing differences between the telencephalon and diencephalon. Lastly, we also showed that fli1a+ mesenchymal cells migrate along the presumptive commissure regions before and during midline axon crossing. Furthermore, following commissure maturation, specific blood vessels formed at the midline of the POC and immediately ventral and parallel to the AC. This comprehensive account of the cellular populations that correlate with the timing and position of commissural axon pathfinding has supported the conceptual modeling and identification of the early forebrain architecture that may be necessary for proper commissure development.