colon cancer visiumhd data  (10X Genomics)

Journal: Genome Biology

Article Title: STHD: probabilistic cell typing of single spots in whole transcriptome spatial data with high definition

doi: 10.1186/s13059-025-03608-4

Figure Lengend Snippet: Spot-level predictions and benchmarking of a zoomed patch in a colon cancer VisiumHD sample. a Example of STHD results of a patch from the human colon cancer VisiumHD sample, revealing tumor-like epithelial cells and intercryptic immune population, visualized using the interactive STHDviewer. b Cell type posterior probabilities at spot level for the same patch, visualized using Squidpy. Top row, three epithelial cell types; Bottom row, two immune cell types. c A spatial plot of unsupervised Leiden clustering on spot gene expression. d Spatial plots of unsupervised Leiden clustering of gene expression aggregated by bins in size of 4 × 4 spots. Left, clustering bins within patch, right, global cluster group of bins. e Cell type proportion decomposed by RCTD on bins of size of 4 × 4 spots, demonstrating two tumor epithelial classes, enterocytes, and stem-like cells. f The H&E histopathology image in full resolution of the same crop. g Comparison to other cell typing and clustering methods adaptive to high-resolution spots. h The STHD predicted labels at spot-level and STHD-guided binning in bins of 4 × 4 spots and 8 × 8 spots, visualized using Squidpy using the same colon cancer cell type colormap. i Benchmarking with simulated high-resolution spatial data. Left, ground truth cell type labels; middle, STHD predicted cell type labels; right, receiver operating characteristic curve (ROC) curve for all cell types. j Benchmarking against other methods for spot annotation in repeated spatial transcriptomics simulations. k Expression dot plots for marker genes of normal epithelial cells, at spot level or STHD-guided bin level. Left to right: normal epithelial cell types in reference human colon cancer sample, normal epithelial cell types based on STHD predicted spots, normal epithelial cell types from STHD-guided bins of size 4 × 4 spots, normal epithelial cell types in STHD-guided bins of size 8 × 8 spots. The same cell type-specific gene markers from the original colon cancer study were used. ROC, Receiver operating characteristic. AUC, area under the curve

Article Snippet: Colon cancer VisiumHD data (patient P1, P3, P5) from Oliveira et al. [ ] include two colon cancer and one normal colon, available at https://www.10xgenomics.com/products/visium-hd-spatial-gene-expression/dataset-human-crc [ ].

Techniques: Gene Expression, Histopathology, Comparison, Expressing, Marker

Journal: Genome Biology

Article Title: STHD: probabilistic cell typing of single spots in whole transcriptome spatial data with high definition

doi: 10.1186/s13059-025-03608-4

Figure Lengend Snippet: STHD prediction facilitates segmentation of global structures and cell type-specific differential analyses across spatial regions. a Interactive spatial visualization by STHDviewer of predicted cell type labels for spots in the entire human colon cancer VisiumHD sample. b Annotation of regions of interest, separating epithelial regions and two T cell-infiltrated regions. c Cell type proportion bar plots at cell lineage and cell type level for epithelial regions and whole map. Upper: Proportions of major cell lineages including epithelial cells, fibroblasts, B cells, myeloid cells, and T cells. Bottom: Proportions of finer cell types of epithelial lineage. d Spot-level gene expression for epithelial-specific differential genes across regions. Top three genes by log fold change were shown. e Pathway enrichment of epithelial cell-specific genes differentially expressed in the Epi-Tumor-top region. f Macrophage-specific differential genes for the Epi-Tumor-inside region, where differential analyses were performed at bin level of size 4 × 4 spots with cell type stratification from STHD. DE, differential expression

Article Snippet: Colon cancer VisiumHD data (patient P1, P3, P5) from Oliveira et al. [ ] include two colon cancer and one normal colon, available at https://www.10xgenomics.com/products/visium-hd-spatial-gene-expression/dataset-human-crc [ ].

Techniques: Gene Expression, Quantitative Proteomics

Journal: Genome Biology

Article Title: STHD: probabilistic cell typing of single spots in whole transcriptome spatial data with high definition

doi: 10.1186/s13059-025-03608-4

Figure Lengend Snippet: STHD is scalable to other samples, revealing subcellular cell type identities and tissue structures. a STHD spot-level inference of a human colon cancer Visium HD sample showing tumor, colon tissue, and enriched lymphoid structure. b Left, spots (2 × 2 µm) labeled as germinal center B cells, with B cell marker expression aggregated per 8 × 8 µm bins. Right, spots (2 × 2 µm) labeled as T cells, and expression of TLS-specific T cell cytokine CXCL13 by 8 × 8 µm bins. Marker expression plots are illustrated by loupe browser. c Spot-level inference of human colon VisiumHD sample P5 (cancer) and P3 (normal adjacent tissue). d Immune cell-specific composition across all four human colon Visium HD samples. e STHD results for the mouse small intestine VisiumHD sample, with two zoomed-in regions highlighted. f Spots with immune cell identities in the lymph node-like region, showing T cell, Ms4a1 + B cell and Ighd + B cells. Top, STHD labels in 2 × 2 µm spots, Bottom, marker gene expression in 8 × 8 µm bins. g Spots of intestine cells forming layered structure, including intestine enterocyte cell, epithelial cell, goblet cell, Paneth cell, and fibroblasts. Top, STHD labels in 2 × 2 µm spots, Bottom, marker gene expression in 8 × 8 µm bins. GC, germinal center. CRC, colorectal cancer. NAT, normal adjacent tissue

Article Snippet: Colon cancer VisiumHD data (patient P1, P3, P5) from Oliveira et al. [ ] include two colon cancer and one normal colon, available at https://www.10xgenomics.com/products/visium-hd-spatial-gene-expression/dataset-human-crc [ ].

Techniques: Labeling, Marker, Expressing, Gene Expression