Journal: bioRxiv

Article Title: Exploiting pair correlation function to describe biological tissue structure

doi: 10.64898/2025.12.19.695425

Figure Lengend Snippet: (A) Analytical workflow. (B) Example of pcf fit. Grey dots correspond to the empirical pcf. Each colored curve corresponds to the fit of a different model. The dashed line corresponds to the line of equation y=1. (C) Distribution of R 2 values for the different model fits to the IMC Lymph node dataset. The thick line corresponds to the median, and the bottom and upper limits of the box correspond to the first and third quartiles, respectively. The lower and upper whiskers correspond to the lowest and highest values, respectively, within the range of the first and third quartiles ±1.5 times the interquartile range (IQR). All boxplots shown in this manuscript are using the same graphical code. The boxplot is based on N = 17 cell types. (D) Distribution of R 2 values for the different model fit to the Xenium Lymph node dataset. The boxplot is based on N = 18 cell types. (E) Distribution of R 2 values for the different model fit to the MERIFSH Tonsil dataset. The boxplot is based on N = 24 cell types. (F) Distribution of R 2 values for the different model fit to the CosMX Prefrontal cortex dataset. The boxplot is based on N = 14 cell types. ( G ) Comparison of the inferred τ parameter with the scale parameter of the simulated Thomas point pattern. Each dot is the average result of 50 simulations. The dashed line corresponds to a linear regression. The black bars correspond to standard deviation observed across simulations. ( H ) Comparison of the inferred C normalised parameter with the noise parameter of the simulated noisy Thomas point pattern. Each dot is the average result of 50 simulations. The dashed line corresponds to a linear regression. The black bars correspond to standard deviation observed across simulations. (I) Spatial pattern of 9 cell types from the MERFISH Tonsild dataset. For each PCF-SiM parameter, 3 patterns are shown with an increasing parameter value from to to bottom.

Article Snippet: The three spatial transcriptomic datasets were obtained using the following procedure: – We obtained the Xenium human lymph node dataset from the 10X Genomics website ( https://www.10xgenomics.com/datasets/human-lymph-node-preview-data-xenium-human-multi-tissue-and-cancer-panel-1-standard ).

Techniques: Comparison, Standard Deviation

Journal: bioRxiv

Article Title: Exploiting pair correlation function to describe biological tissue structure

doi: 10.64898/2025.12.19.695425

Figure Lengend Snippet: (A) Pearson’s correlation between the Clark-Evans index and the three PCF-SiM fitted parameters for the IMC Lymph node dataset. ( B ) Pearson’s correlation between the Clark-Evans index and the three PCF-SiM fitted parameters for the Xenium Lymph node dataset. ( C ) Pearson’s correlation between the Clark-Evans index and the three PCF-SiM fitted parameter for the MERFISH Tonsil dataset. ( D ) Pearson’s correlation between the Clark-Evans index and the three PCF-SiM fitted parameters for the CosMX Prefrontal cortex dataset. (E) Pearson’s correlation between the overdispersion index and the three PCF-SiM fitted parameters for the IMC Lymph node dataset. (F) Pearson’s correlation between the overdispersion index and the three PCF-SiM fitted parameters for the Xenium Lymph node dataset. (G) Pearson’s correlation between the overdispersion index and the three PCF-SiM fitted parameters for the MERFISH Tonsil dataset. (H) Pearson’s correlation between the overdispersion index and the three PCF-SiM fitted parameters for the MERFISH Tonsil dataset. (I) Comparison of the initially estimated C normalised parameter and the estimated C normalised parameter after label corruption. Each dot corresponds to a cell type and is the result of 100 simulations. Dashed line corresponds to the x=y line. The black bars correspond to the standard deviation observed across samples. (J) Comparison of the initially estimated p parameter and the estimated p parameter after label corruption. Each dot corresponds to a cell type and is the result of 100 simulations. Dashed line corresponds to the x=y line. The black bars correspond to the standard deviation observed across samples. (K) Comparison of the initially estimated τ parameter and the estimated τ parameter after label corruption. Each dot corresponds to a cell type and is the result of 100 simulations. Dashed line corresponds to the x=y line. The black bars correspond to the standard deviation observed across samples. ( L ) Estimated τ parameter for the different proteins of the Influenza data. The large horizontal lines correspond to the median while the two smaller horizontal bars correspond to the interquartile ranges. All “prism-like” plots in the manuscript are using the same graphical codes. P-values were computed using Tukey’s honestly significant difference test. ( M ) Estimated p parameter for the M2 protein according to the protein genotype. P-value was computed using a regular Welsh test. ( N ) Estimated PCF-SiM parameters for the different tree species from the Lansing Woods dataset.

Article Snippet: The three spatial transcriptomic datasets were obtained using the following procedure: – We obtained the Xenium human lymph node dataset from the 10X Genomics website ( https://www.10xgenomics.com/datasets/human-lymph-node-preview-data-xenium-human-multi-tissue-and-cancer-panel-1-standard ).

Techniques: Comparison, Standard Deviation

Journal: European Journal of Immunology

Article Title: Identification and Mapping of Human Lymph Node Stromal Cell Subsets by Combining Single‐Cell RNA Sequencing with Spatial Transcriptomics

doi: 10.1002/eji.202451218

Figure Lengend Snippet: Comprehensive single‐cell RNA sequencing strategy to characterize major lymph node stromal cell (LNSC) populations. (a) Overview of the experimental pipeline used in this study. The process begins with the enzymatic digestion of a human lymph node to generate a single‐cell suspension, which is then subjected to cell sorting and sequencing using the 10× Chromium platform. The resulting data is analyzed to produce a UMAP (Uniform Manifold Approximation and Projection) representation, which clusters the cells based on their gene expression profiles. These clusters are then mapped back to their spatial locations within the lymph node tissue using spatial transcriptomics, providing a physical localization of the different fibroblast subsets. Finally, multiplexed imaging microscopy is used to validate the physical location of these subsets within the lymph node using a fluorescent microscope. (b) Flow cytometry gating strategy used to sort viable CD45‐ CD235a‐ single cells containing DN, FRC, LEC, and BEC (see Figure for more details). (c) Two‐dimensional clustering of 9267 LNSCs analyzed by scRNA‐seq. LNSCs are color‐coded based on their main phenotype: fibroblasts (red), LEC (blue), and BEC (green). (d) Pie chart illustrating the proportional distribution of each LNSC phenotype among the total 9267 cells analyzed. This chart provides a quantitative overview of the relative abundance of fibroblasts, LEC, and BEC in the sample. (e) Violin plots showing the log‐normalized expression levels of marker genes specific to each LNSC phenotype. For endothelial cells (BEC and LEC), PECAM1 and VWF are used as markers; for fibroblasts (FRC), DCN, LUM, and PDGFRB are shown; and for LEC, PROX1 and LYVE1 are the defining marker. (f) Heatmap depicting the top 50 differentially expressed genes across the three main LNSC populations at the single‐cell level. Marker genes are highlighted within the heatmap, and the bars at the top of the heatmap correspond to the different LNSC phenotypes (red for fibroblasts, blue for LEC, and green for BEC). Z‐scores are shown to indicate fold‐change expression relative to the mean expression for each gene.

Article Snippet: The human lymph node spatial dataset was downloaded from the dataset portal provided by 10× Genomics ( https://support.10xgenomics.com/spatial‐gene‐expression/datasets/1.1.0/V1_Human_Lymph_Node ).

Techniques: RNA Sequencing, Suspension, FACS, Sequencing, Gene Expression, Imaging, Microscopy, Flow Cytometry, Expressing, Marker

Journal: European Journal of Immunology

Article Title: Identification and Mapping of Human Lymph Node Stromal Cell Subsets by Combining Single‐Cell RNA Sequencing with Spatial Transcriptomics

doi: 10.1002/eji.202451218

Figure Lengend Snippet: Identification and characterization of 10 lymph node fibroblast subsets via single‐cell RNA sequencing. (a) UMAP (Uniform Manifold Approximation and Projection) analysis depicting 10 distinct clusters of lymph node fibroblasts. Each cluster represents a subset of fibroblasts with a unique transcriptional profile. Cells are color‐coded according to the cluster to which they belong. (b) Bar plot displaying the absolute number of cells within each of the 10 fibroblast clusters. Each bar is colored to correspond with the UMAP clusters, providing a quantitative overview of the relative abundance of each fibroblast subset. (c) Heatmap showing per cell (column) the expression patterns of the top 50 DEGs (rows) across the 10 fibroblast clusters. Each column represents a single cell, while rows correspond to specific DEGs. Marker genes defining each fibroblast cluster are highlighted on the left, and the Z‐score normalized expression levels are indicated by the color legend on the right. The top of the heatmap includes color‐coded bars representing the assigned cluster for each fibroblast. (d) Violin plots showing Z‐score expression levels of key marker genes for each fibroblast cluster. (e) Gene Ontology (GO) enrichment analysis was conducted on the DEGs identified in each fibroblast cluster. The analysis highlights the top three GO molecular functions for each subset. (f) Integrated scRNA‐seq analysis of LN fibroblasts by combining our dataset with two publicly available datasets of Abe et al. and Kapoor et al. . UMAP visualization illustrates the distribution of cell subsets across the integrated dataset, comprising 42,073 lymph node cells isolated from a total of 13 donors. The integrated analysis reveals the consistency of the fibroblast clusters across different datasets. (g) This stacked bar plot shows the percentage distribution of cell clusters within three integrated datasets: Abe et al. , Grasso et al., and Kapoor et al. . Each bar represents one dataset ( x ‐axis), and the different colors within each bar correspond to specific clusters (0–11), as indicated in the legend on the right. The y ‐axis represents the fraction of cells within each dataset, ranging from 0.0 to 1.0, with 1.0 indicating 100% of cells within a dataset.

Article Snippet: The human lymph node spatial dataset was downloaded from the dataset portal provided by 10× Genomics ( https://support.10xgenomics.com/spatial‐gene‐expression/datasets/1.1.0/V1_Human_Lymph_Node ).

Techniques: RNA Sequencing, Expressing, Marker, Isolation

Journal: European Journal of Immunology

Article Title: Identification and Mapping of Human Lymph Node Stromal Cell Subsets by Combining Single‐Cell RNA Sequencing with Spatial Transcriptomics

doi: 10.1002/eji.202451218

Figure Lengend Snippet: Spatial distribution of fibroblast subsets within the lymph node revealed by spatial transcriptomics and microscopy. (a) On the left, a histological section of a human lymph node stained with hematoxylin and eosin (H&E), highlighting the structural components of the lymph node, such as follicles, germinal centers, and medullary areas. The image is adapted from 10× Genomics. On the right, spatial gene expression profiles corresponding to the same histological section, reveal distinct regions such as germinal centers, follicles, T‐ and B‐cell areas, medulla, blood vessels, and lymphatic vessels. Each region is color‐coded (see legend) to illustrate its spatial distribution (see also Figure for more details on the analysis). (b) Heatmap showing the differentially expressed genes (DEGs) used to define and annotate the major regions of the lymph node. The color‐coded bars correspond to each region, with Z‐scores on the right representing the relative gene expression levels across these regions. (C–N) Spatial maps displaying in‐silico prediction scores for the presence of each fibroblast subset at various locations within the lymph node. Prediction scores range from 0 (blue, indicating no predicted presence) to 1 (red, indicating a high likelihood of presence). (O) Two‐dimensional heatmap showing the clustering of mean prediction scores per area of the lymph node for each fibroblast subset. The mean prediction values were grouped into low, medium, and high scores, indicating the likely abundance of each fibroblast subset across different regions of the lymph node. (P) Composite immunofluorescence micrographs (acquired at 40× magnification) of a human lymph node section. Specific markers, including CD31 (green), CD34 (magenta), PDPN (yellow), and CD19 (cyan), are displayed in individual panels, and the “Merge” panel combines the staining patterns of CD31, CD34, and PDPN to reveal co‐localization. The “Merge + Sytox Blue” panel adds a nuclear counterstain. Inserts in the micrographs magnify regions of particular interest, with arrows indicating co‐localization of specific markers. Scale bars are 20 µm for individual panels and inserts, and 100 µm for the full section. Representative images from one of three different lymph nodes are shown. (Q) Immunofluorescence staining of another human lymph node section, (acquired at 40× magnification), showing the distribution of markers CD21L (magenta), GLDN (green), PDPN (yellow), and CD19 (cyan) within the tissue architecture. The full section (leftmost panel) illustrates the spatial localization of these markers. The upper and lower rows of magnified inserts focus on two distinct areas of interest within the lymph node. Individual panels provide detailed views of each marker, while the final panel merges all the markers along with Sytox Blue as a nuclear counterstain. Scale bars are 20 µm for individual panels and inserts, and 100 µm for the full section. Representative images from one of three different lymph nodes are shown.

Article Snippet: The human lymph node spatial dataset was downloaded from the dataset portal provided by 10× Genomics ( https://support.10xgenomics.com/spatial‐gene‐expression/datasets/1.1.0/V1_Human_Lymph_Node ).

Techniques: Microscopy, Staining, Gene Expression, In Silico, Immunofluorescence, Marker

Journal: Genome Biology

Article Title: DeepGFT: identifying spatial domains in spatial transcriptomics of complex and 3D tissue using deep learning and graph Fourier transform

doi: 10.1186/s13059-025-03631-5

Figure Lengend Snippet: DeepGFT effectively characterized the subtle tissue architectures of human lymph node Visium data. a H&E image and ground truth segmentation of germinal center. b Domain assignments (evaluated by Jaccard index) and results generated by DeepGFT, BayesSpace, SpaGCN, SpaceFlow, Scanpy, GraphST, STAGATE, stLearn, SpatialPCA, and spaVAE in human lymph node section. c Evaluation results of four metrics (Jaccard index, precision, recall, and F1 score) of ten tools. d The GCs and their adjacent regions obtained from DeepGFT, as well as genes supported corresponding spatial domains. e Cell type compositions of each spatial domain, obtained from cell2location. f Biological processes involved in DEGs in domain 1

Article Snippet: The human lymph node dataset (10x Visium) is available at ( https://www.10xgenomics.com/cn/resources/datasets/human-lymph-node-1-standard-1-1-0 ) [ ].

Techniques: Generated

Journal: APL Bioengineering

Article Title: Interstitial fluid flow in an engineered human lymph node stroma model modulates T cell egress and stromal change

doi: 10.1063/5.0247363

Figure Lengend Snippet: Lymph node stroma model development. We sought to develop a model of immune cell egress from the lymph node (a). A representative immunofluorescence image of a portion of a normal human lymph node is shown. In human lymph nodes, lymphatics (PDPN+, CD31+, visualized in red and green) and fibroblastic reticular cells (PDPN+, in green) form the structure of the T cell zone (b). Schematic of an intended structure of the T cell zone model (c). LECs were seeded on the underside of a tissue culture inset. PhotoHA-collagen gels laden with FRCs are crosslinked above the LECs and then incubated for 30 min (d). After thermal cross-linking, the final gel consists of hyaluronic acid and collagen (e). With this methodology, FRCs formed networks (f) and LECs formed an intact monolayer (g). Scale bars are 50 μ m. Magnetic resonance imaging (MRI) demonstrates altered fluid transport in the presence of LN stroma (h), (i). Divergence of fluid was significantly decreased in the presence of LN stroma (j). Each data point represents a biological replicate (n = 3). Significance was determined by Students' t-test, with significant p values (<0.05) reported on the graph.

Article Snippet: Human lymph node lymphatic endothelial cells (LECs) (Sciencell) were cultured on fibronectin coated flasks in VascuLife® VEGF-Mv Endothelial Complete media, containing 5% fetal bovine serum and additional supplements as provided by the supplier (Lifeline Cell Technology).

Techniques: Immunofluorescence, Incubation, Magnetic Resonance Imaging

Journal: APL Bioengineering

Article Title: Interstitial fluid flow in an engineered human lymph node stroma model modulates T cell egress and stromal change

doi: 10.1063/5.0247363

Figure Lengend Snippet: LEC barrier integrity is decreased under IFF. Pressure heads of media are utilized to drive IFF (a). To create the highest magnitude flow rate, PE50 tubing was secured onto the tissue culture inset with no leaks (b). 3D printed lid raisers to the specifications of a 12 well plate (c) maintain sterility in the incubator. LEC proliferation is unchanged by IFF, as quantified by % of EdU+ cells (d). Representative images show LECs visualized with CD31 (gray). Nuclei are stained with DAPI (blue). Scale bars are 100 μ m (e). LEC coverage (f) and disrupted junctions (g) are quantified. Each point represents a biological replicate from an independent experiment, for n = 3. Each data point represents a biological replicate (n = 3). Significance was determined by two-way ANOVA followed by Tukey's t-test, with significant p values (<0.05) reported on each graph.

Article Snippet: Human lymph node lymphatic endothelial cells (LECs) (Sciencell) were cultured on fibronectin coated flasks in VascuLife® VEGF-Mv Endothelial Complete media, containing 5% fetal bovine serum and additional supplements as provided by the supplier (Lifeline Cell Technology).

Techniques: Sterility, Staining

Journal: APL Bioengineering

Article Title: Interstitial fluid flow in an engineered human lymph node stroma model modulates T cell egress and stromal change

doi: 10.1063/5.0247363

Figure Lengend Snippet: Presence of T cells disrupts LEC junctions regardless of flow and ameliorates flow-induced changes to FRCs. LECs are visualized in representative images with CD31 in gray and DAPI in blue (scale bar 100 μ m), and FRCs are visualized with F-actin in green (scale bar 50 μ m), in the presence of naïve CD8+ T cells (blue) (a). Quantification of LEC monolayer coverage (b) and disrupted junctions (c). FRC coverage in the presence of T cells and IFF (d). Scale bar is 50 μ m. Each data point represents a biological replicate (n = 3). Significance was determined by two-way ANOVA followed by Tukey's t-test, with significant p values (<0.05) reported on each graph.

Article Snippet: Human lymph node lymphatic endothelial cells (LECs) (Sciencell) were cultured on fibronectin coated flasks in VascuLife® VEGF-Mv Endothelial Complete media, containing 5% fetal bovine serum and additional supplements as provided by the supplier (Lifeline Cell Technology).

Techniques: