Journal: Briefings in Bioinformatics

Article Title: Cross-modality representation and multi-sample integration of spatially resolved omics data

doi: 10.1093/bib/bbag214

Figure Lengend Snippet: PRESENT facilitates accurate spatial domain identification in spatial RNA-ADT data. (a) spatial visualization of the 10x Genomics Visium RNA-protein human lymph node sample colored by ground truth domain labels. (b) Quantitative comparison of spatial domain identification performance between PRESENT and other baseline methods, shown as a bar plot for the human lymph node dataset. (c) Quantitative comparison between PRESENT utilizing both RNA and ADT data (RNA & ADT) and PRESENT using only RNA (RNA-only) or ADT data (ADT-only), shown as a radar plot in the human lymph node dataset. (d) UMAP visualization of latent embeddings from different methods, colored by ground truth domain labels in the human lymph node dataset. (e) UMAP visualization of latent embeddings, colored by cluster labels in the human lymph node dataset. (f) Spatial visualization of spots colored by cluster labels in the human lymph node dataset. The cluster labels in e and f were derived from latent embeddings of different methods using the Leiden algorithm. (g) Histology image of the SPOTS mouse spleen dataset and spatial visualization of spots colored by cluster labels in the SPOTS mouse spleen dataset. The cluster labels were derived from latent embeddings of PRESENT using Leiden algorithm. (h) Differentially expressed proteins of each spatial domain through one-versus-all Wilcoxon rank-sum test, shown as a dot plot. (i) DEGs of all the spatial domains through one-versus-all Wilcoxon rank-sum test, shown as a dot plot. (j) Volcano plot showing the DEGs of Mac1-enriched domain and Mac2-enriched domain through Mac1-versus-Mac2 Wilcoxon rank-sum test, where the x axis denotes the log(fold-change) (log(FC)), while the y axis denotes the significance measured by -log10(false discovery rate) (−log10(FDR)). The vertical dashed line represents the threshold for log(FC)= \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} $\pm$\end{document} 0.2, while the horizontal dashed line denotes the threshold for -log10(FDR) = 0.05. (k) Chord plot demonstrating the linkage of DEGs in the Mac1-enriched domain and the corresponding enriched pathways. The left semicircle represents DEGs while the right semicircle denotes the enriched biological processes. Bar plot is employed to demonstrate the significance of each pathway (x axis, −log10(FDR)). (l) The linkage of DEGs in the Mac2-enriched domain and corresponding pathways as well as the significance of each enriched pathway.

Article Snippet: The 10x Visium mouse brain datasets, including a sagittal anterior section and a sagittal posterior section, are accessible at the 10x Genomics websites https://www.10xgenomics.com/datasets/mouse-brain-serial-section-2-sagittal-anterior-1-standard and https://www.10xgenomics.com/datasets/mouse-brain-serial-section-2-sagittal-posterior-1-standard , respectively.

Techniques: Comparison, Derivative Assay

Journal: Briefings in Bioinformatics

Article Title: Cross-modality representation and multi-sample integration of spatially resolved omics data

doi: 10.1093/bib/bbag214

Figure Lengend Snippet: PRESENT integrates single-omics samples of multiple developmental stages or dissected areas. (a) The quantitative evaluation of different integration methods on the three spatial ATAC mouse embryo samples using 14 metrics divided into two categories, namely batch effect removal and biological variance conservation. The category scores of these two aspects were calculated by averaging the metrics within each category. An overall score for each integration method was computed using a 40/60 weighted mean of the category scores for batch effect removal and biological variance conservation. (b) The spatial visualization of spots across the three spatial ATAC mouse embryo samples colored by ground truth spatial domains. (c) The spatial visualization of spots across the three spatial ATAC mouse embryo samples colored by the spatial clusters identified based on different integration methods. The first, second and third row of b and c denotes the samples from E12.5, E13.5 and E15.5 stages, respectively. (d) The anatomic annotation of the sagittal region in P56 mouse brain provided by Allen Reference Atlas . (e) The joint spatial clustering results based on the latent embeddings obtained by STAligner on the two horizontal mouse brain sagittal samples generated by the 10x Genomics Visium platform. (f) The joint spatial clustering results based on the latent embeddings obtained by GraphST on the two horizontal mouse brain sagittal samples generated by the 10x Genomics Visium platform. (g) The joint spatial clustering results based on the latent embeddings obtained by PRESENT on the two horizontal mouse brain sagittal samples generated by the 10x Genomics Visium platform.

Article Snippet: The 10x Visium mouse brain datasets, including a sagittal anterior section and a sagittal posterior section, are accessible at the 10x Genomics websites https://www.10xgenomics.com/datasets/mouse-brain-serial-section-2-sagittal-anterior-1-standard and https://www.10xgenomics.com/datasets/mouse-brain-serial-section-2-sagittal-posterior-1-standard , respectively.

Techniques: Generated

Journal: Nature Communications

Article Title: Tumor cell villages define the co-dependency of tumor and microenvironment in liver cancer

doi: 10.1038/s41467-026-69797-z

Figure Lengend Snippet: A Schematic overview of the workflow in this study. A total of 50 samples collected from the tumor core (T), tumor border (B), and adjacent non-tumor tissue (N) of 7 liver cancer patients (4 HCC, 3 iCCA) were profiled. Sample IDs were named based on histological subtypes of liver cancer, where H represents HCC and C represents iCCA. Single-cell transcriptome data, 10X Visium spatial transcriptome data, and bulk transcriptome data were used for validation, with sample numbers indicated. Illustration was partially created using BioRender. B UMAP embeddings of all profiled single cells colored by cell types ( n = 2,460,095, top panel), and the cells colored by profiling methods ( n = 2,347,589, CosMx TM SMI; n = 112,506, scRNA-seq, bottom panel). C , D A representative tumor sample (1CT) colored by cell types ( C ) and gene score of each cell type ( D ). Gene score was determined based on the average expression of marker genes specific to each cell type. E Cell type annotation and protein staining of a selected window in ( C ). CD68 (red) and Pan-cytokeratin (Pan-CK, green) represent markers for macrophages and epithelial cells. DAPI (light gray) and CD298/B2M (blue) were used for nuclei and membrane staining. A total of 15 samples were profiled using CosMx TM SMI. Scale bars, 50 µm. F Comparison of cell type compositions based on scRNA-seq and CosMx TM SMI from the same set of liver cancer patients in our cohort.

Article Snippet: The publicly available 10X genomics visium datasets used in this study include samples from Liu et al. (Mendeley Data: skrx2fz79n, https://data.mendeley.com/datasets/skrx2fz79n ), Wu et al. ( http://lifeome.net/supp/livercancer-st/data.htm ), Zhang et al. (GEO accession: GSE238264 , https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE238264 ), and Mo et al. (HTAN DCC Portal under the HTAN WUSTL Atlas, https://data.humantumoratlas.org/ ).

Techniques: Single Cell, Biomarker Discovery, Expressing, Marker, Staining, Membrane, Comparison

Journal: Nature Communications

Article Title: Tumor cell villages define the co-dependency of tumor and microenvironment in liver cancer

doi: 10.1038/s41467-026-69797-z

Figure Lengend Snippet: A Enrichment of each tumor cell state on SDNs. Red lines indicate observed enrichment while gray lines stand for randomized enrichment score. P -values were calculated based on the observed and randomized values ( n = 100 iterations per condition, see Methods for details). No adjustment was made for multiple comparisons. **, p -value < 0.01. Source data and exact p -values are provided as a Source data file. B The distances between EMT-like malignant cells and endothelial cells ( n = 20,312), as well as the distances between other malignant cells (cell cycle, cholangiocyte-like, or immune response and locomotion) and endothelial cells ( n = 433,456). Each box shows the median (center line), interquartile range (box), and data range (whiskers). Only tumor cell states with no significant enrichment in the SDN-T1–SDN-T5 were included in the comparison with EMT-like malignant cells. P -value was calculated with a one-sided Student’s t -test. ***, p -value < 0.001. Detailed statistics and source data are provided as a Source Data file. C The number (left) and the proportion (right) of malignant cells in the surrounding (within 40 µm distance) of cell cycle-related malignant cells ( n = 125,867) compared to that of other malignant cells ( n = 665,674). Each box shows the median (center line), interquartile range (box), and data range (whiskers). P -value was calculated using one-sided Student’s t -test. ***, p -value < 0.001. Detailed statistics and source data are provided as a Source Data file. D Validation of the spatial preference of tumor cell states using 10X Visium data from Liu et al. (left), Wu et al. (middle), and Zhang et al. (right). Each pair of connected dot and triangle represents a tumor cell state. For each tumor cell state, “Hit” represents the significantly associated SDNs in (A), while “non-Hit” indicates the rest of the SDNs ( n = 12 tumor cell states). Violin plots of the distributions of the enrichment scores were shown. p -value was calculated with a one-sided paired Student’s t -test. ***, p -value < 0.001. Source data and detailed statistics are provided as a Source Data file.

Article Snippet: The publicly available 10X genomics visium datasets used in this study include samples from Liu et al. (Mendeley Data: skrx2fz79n, https://data.mendeley.com/datasets/skrx2fz79n ), Wu et al. ( http://lifeome.net/supp/livercancer-st/data.htm ), Zhang et al. (GEO accession: GSE238264 , https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE238264 ), and Mo et al. (HTAN DCC Portal under the HTAN WUSTL Atlas, https://data.humantumoratlas.org/ ).

Techniques: Comparison, Biomarker Discovery

Journal: Communications Biology

Article Title: Advancing spatial cellular communication inference with ligand diffusion and transport model

doi: 10.1038/s42003-025-09413-w

Figure Lengend Snippet: a Spatial annotation of the dataset, including labeled regions (left), aggregated compartments (middle), and the matched H&E image (right). Each spot has a diameter of approximately 55 μm, as specified by 10x Genomics. b Top row: spatial expression of the ligand TGFB1 and its cognate receptors TGFBR2 , TGFBR1 and ACVR1B . Bottom row: total received signals predicted by SCILD for L-R pairs TGFB1- > TGFBR2, TGFB1- > TGFBR1, and TGFB1- > ACVR1B. c Same as ( b ) for the receptor ERBB2 and its cognate ligands HBEGF , AREG , and TGFA .

Article Snippet: The 10x Visium Human Breast Cancer dataset was downloaded from https://www.10xgenomics.com/datasets/human-breast-cancer-block-a-section-1-1-standard-1-1-0 and denoised using Sprod ( https://github.com/yunguan-wang/SPROD ) .

Techniques: Labeling, Expressing