Journal: bioRxiv

Article Title: SpatialBench : Comparative cross-platform benchmarking of high-resolution spatial transcriptomics using matched mouse lymphoid tissue

doi: 10.64898/2026.04.29.721531

Figure Lengend Snippet: A . Schematic overview of the three spatial transcriptomics (ST) platforms used in this study, comprising sequencing-based (sST) and imaging-based (iST) approaches. Visium HD (10x Genomics) is a sST platform that captures spatially barcoded polyadenylated transcripts, followed by cDNA library preparation and sequencing. MERSCOPE (Vizgen) and Xenium (10x Genomics) are iST platforms for in situ transcript detection using multiplexed fluorescence imaging, with MERSCOPE relying on direct fluorescent probe hybridization and Xenium on padlock probes with rolling circle amplification, followed by iterative imaging, signal removal, and rehybridization cycles for transcript decoding. B . Experimental design of the SpatialBench study using spleen sections from malaria infected and cured mice across wild-type (WT), control (CTRL) ( Tbx21 +/+ Cd23 Cre ), and B cell-specific conditional Tbx21 knockout (KO) ( Tbx21 fl/fl Cd23 Cre ) genotypes. Fresh-frozen spleen sections were profiled using Visium HD, MERSCOPE, and Xenium. Matched single-nuclei RNA seq (snRNA-seq) datasets were generated from dissociated formalin-fixed paraffin-embedded (FFPE) WT spleens. C . Sample composition of the SpatialBench dataset. Colors denote platform identity, with MERSCOPE (magenta), Xenium (blue), Visium HD (green), and snRNA-seq (yellow). Asterisk denote Xenium technical replicates with multimodal staining. D . Summary of sample number, spatial resolution, and Venn diagram displaying shared genes between panels of the ST platforms. E . Dataset-level statistics for binned ST data, with 8 µm and 16 µm bins shown in the top and bottom rows, respectively. Columns show total bins, total transcript counts, transcript counts for the subset of 90 shared genes, sparsity, and sparsity for the shared genes across platforms. Points indicate individual samples, with bars and error bars denoting the mean and standard error of the mean (SEM). F . Global Pearson correlation of gene expression between snRNA-seq and ST platforms. Density contour plots show gene-level correlations for Visium HD (green, left), MERSCOPE (magenta, middle), and Xenium (blue, right) relative to snRNA-seq. Gene counts were averaged across WT samples and log 10 -transformed counts per million (CPM). Density contours represent the distribution of genes, with increased color intensity indicating higher density. The diagonal line (slope = 1) denotes equality, and Pearson correlation coefficients (R) and the number of shared genes (n) between ST platforms and snRNA-seq are reported for each comparison. G . Computational workflow for SpatialBench , showing preprocessing, quality control, normalization, dimensionality reduction, and downstream analyses, with platforms color-coded across steps.

Article Snippet: All slide incubation steps were performed by positioning a Visium Low Profile Thermocycler Adapter (10x Genomics, PN-3000823) on a thermal cycler (Bio-Rad C1000 Touch) to ensure efficient and even heat transfer.

Techniques: Spatial Transcriptomics, Sequencing, Imaging, cDNA Library Assay, In Situ, Fluorescence, Hybridization, Amplification, Infection, Control, Knock-Out, RNA Sequencing, Generated, Formalin-fixed Paraffin-Embedded, Staining, Gene Expression, Transformation Assay, Comparison

Journal: bioRxiv

Article Title: SpatialBench : Comparative cross-platform benchmarking of high-resolution spatial transcriptomics using matched mouse lymphoid tissue

doi: 10.64898/2026.04.29.721531

Figure Lengend Snippet: A . Matched spleen sections from wild-type (WT709, WT713) and conditional Tbx21 knockout (KO167, KO168) mice profiled using Visium HD (green), MERSCOPE (magenta), and Xenium (blue). Heatmaps display total transcript counts per 8 µm bin, scaled per platform. Scale bars, 1 mm. B . Bin-level (8 µm) evaluation of median transcript counts per bin (top) and median detected genes per bin (bottom) for the 90 commonly shared genes across Visium HD, MERSCOPE, and Xenium plaforms. Points indicate individual samples. C . Bin-level (8 µm) evaluation of median transcript counts per bin (top) and median detected genes per bin (bottom) for the 14,999 commonly shared genes between Visium HD and snRNA-seq assays. Points indicate individual samples. D . Rarefaction curves showing the expected numbers of detected genes (top) and transcripts (Unique Molecular Identifiers; UMI) (bottom) in downsampled data as a function of mapped reads. M, millions. E . Total gene targeting and negative control probe counts per sample in MERSCOPE and Xenium. Box plots show sample-level totals ( n = 9 per platform). Points indicate individual samples. F . Distributions of observed Moran’s I values for gene targeting probes and background probes (Blanks in MERSCOPE; Unassigned in Xenium) across all genes and samples, summarized by box plots. G . Background to signal ratio (BSR) of background probe detection across MERSCOPE and Xenium. Data are presented as mean ± SEM. H . Rank-ordered mean probe counts (log 10 (mean count + 1)) for MERSCOPE and Xenium across all samples. Shaded area marks the platform-specific 95th percentile of background probe counts. Highlighted points indicate gene targeting probes at or below this threshold. I . Multidimensional scaling (MDS) of pseudo-bulk expression profiles for the 90 common genes across platforms. Point color denotes platform and shape denotes sample group. J . Variance partitioning of pseudo-bulk gene expression profiles attributed to platform (75.8%), biological group (12.1%), and residual (10.7%) variation. K . Cross-platform gene-level correlations. Density contour plots show pairwise comparisons of averaged log 10 CPM expression across shared genes in matched Visium HD, MERSCOPE, and Xenium samples, with Pearson correlation coefficients (R). Points indicate individual genes. Source data are provided as a Source Data file.

Article Snippet: All slide incubation steps were performed by positioning a Visium Low Profile Thermocycler Adapter (10x Genomics, PN-3000823) on a thermal cycler (Bio-Rad C1000 Touch) to ensure efficient and even heat transfer.

Techniques: Knock-Out, Negative Control, Expressing, Gene Expression

Journal: bioRxiv

Article Title: SpatialBench : Comparative cross-platform benchmarking of high-resolution spatial transcriptomics using matched mouse lymphoid tissue

doi: 10.64898/2026.04.29.721531

Figure Lengend Snippet: A . UMAP of snRNA-seq, Visium HD (8 µm bins), and cell-level MERSCOPE and Xenium data showing major spleen cell type annotations used for downstream analyses. B . Bar plot of average silhouette width (ASW) per platform, computed from the Principal component analysis (PCA) embeddings underlying the UMAP shown in . C . Spatial maps of annotated cell types for WT713 and KO167 spleen sections across matched Visium HD (top), MERSCOPE (middle), and Xenium (bottom) samples. Each color represent a cell type. Scale bars, 1 mm. D . Bar plot of cell type proportions in WT and KO samples by platform. Visium HD includes WT and KO samples ( n = 2 each), and MERSCOPE and Xenium include WT and KO samples ( n = 3 each) (top). Data are presented as mean ± SEM. E . Box plot of extracted GC B cell proportions from 4D in WT and KO samples by platform. Each point indicate individual sample. F . Spatial maps of Tbx21 expression in WT713 and KO167 samples measured by Visium HD, MERSCOPE, and Xenium, with expression scaled within each platform. Scale bars, 1 mm. G . Bar plot of the percentage of Tbx21 -positive units in WT and KO samples across platforms, defined as 8 µm bins for Visium HD and cells for MERSCOPE and Xenium. Data are presented as mean ± SEM. H . Bar plot showing the cell type composition of Tbx21 -positive units across platforms, colored by annotated cell type as in panel 4C. I . Volcano plots of pseudo-bulk KO versus WT differential expression in näive B cells (top) and germinal center B cells (bottom) across Visium HD, MERSCOPE, and Xenium. Genes are colored by differential expression status (|log 2 FC| > 1, adjusted P < 0.05). Tbx21 is highlighted in yellow with other key expected down-regulated genes. J . Bar plot of the proportion of significantly differentially expressed genes (adjusted P < 0.05) overlapping an external Tbx21 (encodes for T-bet) knockout gene signature from Ly et al. in näive B and germinal center B cells across platforms. K . Dot plot of pseudo-bulk Tbx21 differential expression across annotated cell types for Visium HD, MERSCOPE, and Xenium. Dot color show log 2 fold change (KO versus WT), with adjusted P -values indicated by dot size and annotated significance indicated by star symbol.

Article Snippet: All slide incubation steps were performed by positioning a Visium Low Profile Thermocycler Adapter (10x Genomics, PN-3000823) on a thermal cycler (Bio-Rad C1000 Touch) to ensure efficient and even heat transfer.

Techniques: Expressing, Quantitative Proteomics, Knock-Out

Journal: bioRxiv

Article Title: SpatialBench : Comparative cross-platform benchmarking of high-resolution spatial transcriptomics using matched mouse lymphoid tissue

doi: 10.64898/2026.04.29.721531

Figure Lengend Snippet: A . Spatial maps of annotated cell types for WT713 and KO167 spleen samples profiled by Visium HD, MERSCOPE, and Xenium. Germinal centers (GC) are delineated with dark and light zones annotated, and representative marker gene expression is shown for Bcl6 (GC), Aicda (dark zone), and Cd83 (light zone). Dotted outlines indicate GC boundaries. Color intensity indicates log-normalised expression. Scale bars, 1 mm (overview) and 200 µm (ROIs). B . Dot plots of dark zone and light zone marker gene expression across platforms. Average expression is shown by color and the percentage of expressing units is indicated by dot size. Visium HD and Xenium show expression stratified by annotated dark and light zones, while MERSCOPE shows expression within GC B cells. C . Bar plot of GC composition in WT and KO samples for Visium HD, MERSCOPE, and Xenium, showing the relative contribution of dark zone, light zone, and unassigned GC regions. Bars represent mean proportions across samples, with colors indicating GC sub-compartments. D-E . Percentage change in ( D ) GC density and ( E ) GC diameter in KO relative to WT samples across platforms. F . Venn diagram showing the overlap of differentially expressed genes (adjusted P < 0.05) between dark and light zones in WT samples and a published GC zone gene set from Victora et al. . G . Volcano plot of pseudo-bulk differential expression between dark and light zones in WT samples profiled with Visium HD. Genes are colored by differential expression status using | log 2 FC| > 1 and adjusted P < 0.05. H .Gene Ontology (GO) biological process enrichment for genes differentially expressed between dark and light zones in WT samples. Bar plots show significantly enriched GO terms for dark zone (top) and light zone (bottom) gene sets, ranked by enrichment significance (log 10 P ). Representative contributing genes are shown for each term. Source data are provided as a Source Data file.

Article Snippet: All slide incubation steps were performed by positioning a Visium Low Profile Thermocycler Adapter (10x Genomics, PN-3000823) on a thermal cycler (Bio-Rad C1000 Touch) to ensure efficient and even heat transfer.

Techniques: Marker, Gene Expression, Expressing, Quantitative Proteomics