Journal: The Journal of Experimental Medicine

Article Title: Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

doi: 10.1084/jem.20251067

Figure Lengend Snippet: Overview of the single-cell and spatial data generated from TB-diseased and control lungs. (A) Schematic showing the experimental flow for the isolation of cells from human lung tissues, generation of single-cell libraries using Seq-Well S 3 . Four TB-negative and nine TB-positive lung samples were processed through scRNA-seq. Shown adjacent to the process flow is a low-dimensional embedding (UMAP) of the 19,632 cells passing quality control annotated with high-level cell types (middle) or detailed cell subtype (right). (B) 10x Visium platform workflow for spatial transcriptomics profiling on FFPE samples from TB-diseased lung resections. 21 of these samples come from current TB patients with detectable M.tb ; 9 came from post-TB patient, where bacteria are no longer detected in BAL TB culture after infection. Samples contain either granulomas, iBALTs, or lung LNs, representing different pathological states.

Article Snippet: Spatial transcriptomics (Visium) samples: A section of lung was cut and transferred to 10% buffered formalin to fix for 24 h, then transferred to 70% ethanol until wax embedding.

Techniques: Generated, Control, Isolation, Bacteria, Infection

Journal: The Journal of Experimental Medicine

Article Title: Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

doi: 10.1084/jem.20251067

Figure Lengend Snippet: Spatial transcriptomics on TB-infected human lung samples and single-cell deconvolution. (A) H&E staining on all 30 lung samples from patients previously infected with TB. Scale bars: 800 μm. Identical images for pid_0037, pid_177, pid_0186, pid_187, pid_0192, pid_199, pid_0209, and pid_304. (B) Examples of manual annotation on granuloma structures on H&E staining images. Scale bars: 800 μm.

Article Snippet: Spatial transcriptomics (Visium) samples: A section of lung was cut and transferred to 10% buffered formalin to fix for 24 h, then transferred to 70% ethanol until wax embedding.

Techniques: Infection, Staining

Journal: The Journal of Experimental Medicine

Article Title: Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

doi: 10.1084/jem.20251067

Figure Lengend Snippet: Single-cell transcriptomic reveals heterogeneity within neutrophil populations with disease-specific difference. (A) Neutrophil ( n = 2,963) subclustering reveals three subclusters (left), also colored by patient ID (middle) and disease condition (right). (B) Volcano plot of differential gene expression results of each neutrophil subcluster compared with the rest. Y axis shows −log10 (BH-adjusted P value); x axis shows log2 fold change between cells in subcluster and outside the subcluster. (C) Heatmap of subtype top 10 differentially expressed (DE) genes in each of the neutrophil subcluster. (D) Expression of marker genes in neutrophil subclusters by disease conditions. (E) Fisher’s exact test on abundance of detailed neutrophil subclusters between TB conditions. Statistical annotations: fold-change >2 (ΔΔ). (F) Cell2loc imputed neutrophil abundance distribution on the Visium dataset grouped by TB and granuloma status (Materials and methods). The 5% quantile of the estimated posterior distribution of cell abundance at each Visium spot is displayed, representing the value of cell abundance that the model has high confidence in. Two-sided Mann–Whitney U test without correction were used for statistical testing. ****: P < 0.0001.

Article Snippet: Spatial transcriptomics (Visium) samples: A section of lung was cut and transferred to 10% buffered formalin to fix for 24 h, then transferred to 70% ethanol until wax embedding.

Techniques: Gene Expression, Expressing, Marker, MANN-WHITNEY

Journal: The Journal of Experimental Medicine

Article Title: Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

doi: 10.1084/jem.20251067

Figure Lengend Snippet: Single-cell transcriptomic reveals heterogeneity within monocyte and macrophage populations with disease-specific difference. (A) Monocyte/macrophage ( n = 8,318) subclustering reveals 10 subclusters (left), also colored by patient ID (middle) and disease condition (right). (B) Heatmap of subtype top 10 DE genes in each of the monocyte/macrophage subcluster. (C) Expression of marker genes in monocyte/macrophage subclusters by disease conditions. (D) Two-sided Fisher’s exact test on abundance of detailed macrophage (left) and monocyte (right) subclusters between TB conditions. Holm’s method was applied to adjust P values for multiple-testing correction. Statistical annotations: P value < 0.05 (*), P value < 0.01 (**), P value < 0.001 (***), fold-change >1 (Δ), fold-change >2 (ΔΔ), and fold-change <1 (∇). (E) Cell2loc imputed macrophage (left) and monocyte (right) abundance distribution on the Visium dataset grouped by TB and granuloma status (Materials and methods). The 5% quantile of the estimated posterior distribution of cell abundance at each Visium spot is displayed, representing the value of cell abundance that the model has high confidence in. Two-sided Mann–Whitney U test without correction were used for statistical testing. Statistical annotations: P value < 0.0001 (****). (F) Similar to E, but grouped by TB status and HIV status.

Article Snippet: Spatial transcriptomics (Visium) samples: A section of lung was cut and transferred to 10% buffered formalin to fix for 24 h, then transferred to 70% ethanol until wax embedding.

Techniques: Expressing, Marker, MANN-WHITNEY

Journal: The Journal of Experimental Medicine

Article Title: Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

doi: 10.1084/jem.20251067

Figure Lengend Snippet: Deconvolution of bulk human LN dataset and fibroblast in spatial and single-cell dataset. (A) Dot plot showing distribution of cell type proportion from deconvolution results on each bulk RNA-seq human LN TB granuloma sample, separated by cell type and colored by TB conditions. Only cell types with significant difference between TB conditions are shown. Two-sided T test with Bonferroni correction was used to compare the means. Statistical annotations: P value < 0.05 (*) and P value < 0.01 (**). (B) Cell2loc imputed fibroblast abundance distribution on the Visium dataset group by TB and granuloma status (Materials and methods). The 5% quantile of the estimated posterior distribution of cell abundance per Visium spot is displayed, representing the value of cell abundance that the model has high confidence in. Two-sided Mann–Whitney U test without correction were used for statistical testing. P value < 0.0001 (****); P value > 0.05 (ns). (C) Same as B, but grouped by HIV and TB status. (D) Bar plot of patient distribution in each fibroblast subcluster. (E) UMAP embedding of fibroblasts colored by HIV status of the sample.

Article Snippet: Spatial transcriptomics (Visium) samples: A section of lung was cut and transferred to 10% buffered formalin to fix for 24 h, then transferred to 70% ethanol until wax embedding.

Techniques: RNA Sequencing, MANN-WHITNEY

Journal: The Journal of Experimental Medicine

Article Title: Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology

doi: 10.1084/jem.20251067

Figure Lengend Snippet: Spatial transcriptomics analysis on post- and current TB lung resections. (A) Heatmap showing the expression of human TB-myofibroblast gene signature and SPP1 + CHI3L1 + macrophage markers on selective tissue slides from patients who are post-TB (top) or current TB (bottom), alongside paired H&E staining (these H&E stains are also shown in together with those other samples used for spatial transcriptomics not shown here). (B) Distribution of human TB-myofibroblast signature expression on the spatial cohort. HIV statuses are shown in different shades of blue for positive or negative. Two-sided Mann–Whitney U test without correction was used for statistical testing. Statistical annotation: P value < 0.0001 (****). (C) Distribution of SPP1 + CHI3L1 + macrophage markers and human TB-myofibroblast signature on the spatial data across all Visium spots. Left two panels: Manual segmentation of the granuloma structure was done to allow separation of the Visium slide into three different regions: in granuloma, on granuloma border (cuff), and outside of granuloma (Materials and methods). Right two panels: The same as left panels with the exception that “on border” = True means on granuloma cuff and False means the rest. Two-sided Mann–Whitney U test without correction was used for statistical testing. Statistical annotation: P value < 0.0001 (****). (D) Correlation between human TB-myofibroblast signature and all macrophage subpopulations’ markers. Each circle represents a Visium sample. Boxplot of the Pearson’s r distribution is shown for each macrophage subtype. Mann–Whitney U test without correction were used for statistical testing. Statistical annotation: P value < 0.0001 (****). (E) Spatially informed ligand–receptor (L–R) analysis using LIANA+ on Visium samples. Examples are shown where SPP1(L)–CD44(R) interactions are being nominated as top L–R pairs. H&E overlaid with pathology annotation for granuloma structures are shown next to heatmap of L–R interaction scores, which are calculated at each Visium spot using spatially weighted Cosine similarity (Materials and methods).

Article Snippet: Spatial transcriptomics (Visium) samples: A section of lung was cut and transferred to 10% buffered formalin to fix for 24 h, then transferred to 70% ethanol until wax embedding.

Techniques: Expressing, Staining, MANN-WHITNEY

Journal: Nature Communications

Article Title: SpaIM: single-cell spatial transcriptomics imputation via style transfer

doi: 10.1038/s41467-025-63185-9

Figure Lengend Snippet: SpaIM comprises an ST autoencoder and an ST generator. Both the ST autoencoder and the ST generator are built on the multilayer recursive style transfer (ReST) layers.

Article Snippet: Spatial transcriptomics (ST) autoencoder The ST autoencoder (Fig. ) comprises multilayer ReST encoders.

Techniques:

Journal: Nature Communications

Article Title: SpaIM: single-cell spatial transcriptomics imputation via style transfer

doi: 10.1038/s41467-025-63185-9

Figure Lengend Snippet: a Benchmarking results on the NanoString CosMx spatial transcriptomics dataset (Lung5–rep3), using evaluation metrics including structural similarity index measure (SSIM) and Jaccard similarity (JS). Data are presented as mean values ± 95% confidence intervals across predicted genes ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$n$$\end{document} n = 2,038). b Spatial visualization of cell types in the whole slide. c Spatial visualization of cell types in specific field of views (FOVs).

Article Snippet: Spatial transcriptomics (ST) autoencoder The ST autoencoder (Fig. ) comprises multilayer ReST encoders.

Techniques:

Journal: Frontiers in Cell and Developmental Biology

Article Title: Acetylcholine in the gingival epithelium drives the pathogenesis of periodontitis

doi: 10.3389/fcell.2025.1701252

Figure Lengend Snippet: Acetylcholine signaling and receptor distribution in the periodontal epithelium. (A) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment of epithelial neurotransmitter signaling-related DEGs. Left: upregulated pathways in PD vs. HC; right: downregulated pathways in TP vs. PD. (B) Expression profiles of acetylcholine receptors across different cell types. Receptors not detected in any cells were excluded. (C) Transcript counts of acetylcholine receptor genes from HOKs RNA-seq data. Five receptors detected in all 12 samples (n = 3 per group), with colored shapes representing groups and black bars indicating the mean. Receptors with low expression (zero counts in some samples) were excluded. (D) Expression of acetylcholine receptors CHRNB1 , CHRNA5 , and CHRNA7 in gingival spatial transcriptomics data.

Article Snippet: Utilizing single-cell RNA sequencing (scRNA-seq) data (205,334 cells, 40 human gingival samples) and gingival spatial transcriptomics data (46,230–25 μm 2 spots), we revealed that the gingival epithelium exhibits the most significant functional reprogramming of neural signaling pathways in the periodontitis state.

Techniques: Expressing, RNA Sequencing

Journal: Frontiers in Cell and Developmental Biology

Article Title: Acetylcholine in the gingival epithelium drives the pathogenesis of periodontitis

doi: 10.3389/fcell.2025.1701252

Figure Lengend Snippet: Distribution of tight junction genes in the periodontal gingival epithelium and their regulation by acetylcholine. (A) Expressions of OCLN, CLDN1, and CDH1 in spatial transcriptomics data. (B) Differential comparison of OCLN, CLDN1, and CDH1 in the epithelial subpopulation of HC versus PD groups (***, p < 0.001). (C) Heatmap showing the expressions of OCLN, CLDN1, and CDH1 for each cluster in HC and PD gingiva. (D) Quantitative polymerase chain reaction (qPCR) validation in HOKs. Data are presented as the mean ± standard error of the mean (SEM). ***, p < 0.001.

Article Snippet: Utilizing single-cell RNA sequencing (scRNA-seq) data (205,334 cells, 40 human gingival samples) and gingival spatial transcriptomics data (46,230–25 μm 2 spots), we revealed that the gingival epithelium exhibits the most significant functional reprogramming of neural signaling pathways in the periodontitis state.

Techniques: Comparison, Real-time Polymerase Chain Reaction, Biomarker Discovery

Journal: Nature Communications

Article Title: SpaIM: single-cell spatial transcriptomics imputation via style transfer

doi: 10.1038/s41467-025-63185-9

Figure Lengend Snippet: a Benchmarking results on the NanoString CosMx spatial transcriptomics dataset (Lung5–rep3), using evaluation metrics including structural similarity index measure (SSIM) and Jaccard similarity (JS). Data are presented as mean values ± 95% confidence intervals across predicted genes ( \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$n$$\end{document} n = 2,038). b Spatial visualization of cell types in the whole slide. c Spatial visualization of cell types in specific field of views (FOVs).

Article Snippet: Spatial transcriptomics (ST) generator A similar architecture (Fig. ) is used to generate ST data from SC data.

Techniques: