Cellular & Molecular Imaging

Over the recent years, fluorescence microscopy has seen substantial progress, transitioning from primarily qualitative analysis to a profoundly quantitative methodology, yet limitations in imaging depth and data quality persist due to tissue opacity. Tissue optical clearing (TOC) methodologies were implemented to overcome this hurdle, enabling the imaging and quantitative analysis of large tissue samples, such, e.g., complete murine organs. Further advancements such as Expansion Microscopy (ExM) merges TOC benefits with improved imaging resolution, converting cells and tissues into tissue-gel hybrids. Yet, its intricate protocols lead to a remarkable, ~10-20× tissue volume expansion, making the acquisition of even thin sections exceedingly challenging in terms of time, labor, and data handling. Here we present a novel solution, ISEE (Improved Segmentation by gEntle Expansion), for both TOC and gentle (0.6-0.7×) tissue expansion for significantly improved precision of fluorescent signal segmentation, while maintaining tolerable increase during data acquisition in everyday laboratory practice. When applied to thick ~40-200 µm tissue slices, ISEE renders them transparent and ultimately expanded within minutes, in a simple, one-step process compatible with immunohistochemistry as well as with all major murine organs.

Correlative live cell imaging of landmark organelles — such as nuclei, nucleoli, cell membranes, nuclear envelope and lipid droplets — is critical for systems cell biology and drug discovery. However, achieving this with molecular labels is challenging. Virtual staining of multiple organelles and cell states from label-free images with deep neural networks is an emerging solution. This approach frees the light spectrum for imaging molecular sensors, photomanipulation, or other tasks. Current methods for virtual staining of landmark organelles often fail in the presence of nuisance variations in imaging, culture conditions, and cell types. We report training protocols and a flexible convolutional architecture, UNeXt2, that enable robust virtual staining of nuclei and membranes across diverse imaging parameters, cell states, and types. The strategies include self-supervised and supervised pre-training, improving robustness for multiple cell types — including human cell lines, zebrafish neuromasts, stem cells (iPSCs), and iPSC-derived neurons (iNeurons) — under a range of imaging conditions. We assess models using intensity, segmentation, and application-specific measurements obtained from virtually and experimentally stained nuclei and membranes. These models rescue missing labels, correct non-uniform expression, and mitigate photobleaching. We share three pre-trained models and a PyTorch-based pipeline (VisCy) for training, inference, and deployment.

The Darwin Tree of Life project has the goal to sequence the genomes of 70,000 species of eukaryotic organisms in Britain and Ireland. This is a collection of photographs of the samples included in the study, provided by the National History Museum (NHM).

The European Reference Genome Atlas (ERGA) initiative is a pan-European scientific response to current threats to biodiversity. Reference genomes provide the most complete insight into the genetic basis that forms each species and represent a powerful resource in understanding how biodiversity functions. This is a collection of the samples included in the study, provided by COPO at Earlham Institute.

Files for accuracy measures / ultramicrotome calibration

raw image files analyzed for publication; source data files

Instance segmentation masks for individual arabidopsis roots

Morphological microglia cell types were characterized by image analysis of primary cultures of mixed neurons and glia, in which microglia expresses the tdTomato protein.

Background/Aim: Resistance to chemotherapy is a major obstacle for patients with unresectable colorectal cancer (CRC); however, the factors that induce chemoresistance have not been elucidated. Lipid composition influences neoplastic behaviour. Therefore, this study examined whether lipid composition affects sensitivity to chemotherapeutic agents in CRC. Materials and Methods: We performed a lipidomic analysis of a CRC xenograft-derived spheroid model to identify potential relationships between the lipid profile and chemoresistance to 5-fluorouracil (5-FU). Genetic and pharmacological modulation of lipid synthesis were also used in the HCT-116 and DLD-1 CRC cell lines to further characterize resistance to 5-FU. Results: Our lipidomic profiling revealed that phospholipids with saturated fatty acids (SFAs) were more abundant in 5-FU resistant spheroids. The importance of phospholipids containing SFA in chemoresistance was confirmed by showing that in HCT-116 and DLD-1 cells, genetic or pharmacological inactivation of stearoyl-CoA desaturase-1, a key enzyme that converts SFAs to monounsaturated fatty acids, increased the proportion of SFAs in membranous phospholipids and reduced cell membrane fluidity, and this ultimately resulted in resistance to 5-FU. Conclusion: These data suggest that the saturated to monounsaturated fatty acid ratio in cellular membranous phospholipids affects sensitivity to chemotherapeutic agents.

In Drosophila male gonads, niche-derived Decapentaplegic (Dpp) has a role in maintaining stem cells in close proximity but the range of Dpp diffusion and its role has never been tested in the differentiating cells spaced one-cell layer away. Here, using genetically encoded nanobodies called Morphotraps, we physically block Dpp diffusion without interfering with niche-stem cell signaling and find that a diffusible fraction of Dpp is required to ensure differentiation of GSC daughter cells, opposite of its role in maintenance of GSC in the niche. Our work provides an example in which a soluble niche ligand induces opposed cellular responses in stem cells versus differentiating descendants to ensure spatial control of the niche, which may be a common mechanism to regulate tissue homeostasis.