Journal: Immunity

Article Title: The cytokine TNF promotes transcription factor SREBP activity and binding to inflammatory genes to activate macrophages and limit tissue repair

doi: 10.1016/j.immuni.2019.06.005

Figure Lengend Snippet: KEY RESOURCES TABLE

Article Snippet: SureCell WTA 3′ Library Prep Kit for the ddSeq System , Illumina , Cat# 200142780.

Techniques: Purification, Control, Virus, Plasmid Preparation, Recombinant, Amplex Red Cholesterol Assay, cDNA Synthesis, SYBR Green Assay, Multiplex Assay, RNA Library Preparation, Microarray, Software

Journal: bioRxiv

Article Title: Single-cell transcriptomic atlas of mouse oocyte development from growth to ovulation

doi: 10.64898/2026.03.11.710939

Figure Lengend Snippet: a , UMAP of GCs, colored by seven GC subtypes (Progenitor, Preantral 1, Preantral 2, Mitotic 1, Mitotic 2, Antral Mural and Atretic). b , Feature plots of representative subtype markers on the UMAP. c , Monocle3 pseudotime trajectory inferred for GCs, with the principal graph overlaid and direction indicated from progenitor toward antral mural cells. d , Heatmap of representative genes showing coordinated expression changes along the progenitor-to-mural trajectory (expression shown as z-scores). e , Heatmap of Hallmark ssGSEA scores across granulosa subtypes (z-scored per gene set). f , H&E image of a Stereo-seq FF V1.3 mouse ovary section (6-8 weeks old), with representative regions (α–θ) indicated. Scale bar, 100 μm. g , Cell2location-based spatial mapping of GC subtypes at cell-bin resolution. h , Zoom-in views of representative regions (α, β, γ, and θ) showing H&E morphology and spatial expression of selected marker genes. Scale bar, 50 μm.

Article Snippet: Publicly available mouse ovary Stereo-seq Transcriptomics FF v1.3 demo data were obtained from the STOmics website ( https://www.stomics.tech/col1347 ).

Techniques: Expressing, Marker

Journal: bioRxiv

Article Title: Single-cell transcriptomic atlas of mouse oocyte development from growth to ovulation

doi: 10.64898/2026.03.11.710939

Figure Lengend Snippet: Spatial maps showing cell2location-predicted localization of each annotated cell subtype on the Stereo-seq FF V1.3 ovary section (6–8 weeks old), displayed separately by subtype. Scale bar, 100 μm

Article Snippet: Publicly available mouse ovary Stereo-seq Transcriptomics FF v1.3 demo data were obtained from the STOmics website ( https://www.stomics.tech/col1347 ).

Techniques:

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: a , Inputs to CalicoST are transcript counts X 0 , allele counts Y 0 and D 0 , spatial coordinates S from one or more SRT slices or a 3D alignment of slices. b , CalicoST phases input alleles in Y 0 and D 0 using a database of haplotypes. Optionally, CalicoST infers tumor proportion per spot using the BAF. CalicoST jointly models transcript counts and allele counts as functions of allele-specific copy number states within each clone. CalicoST uses an HMM to model correlations between copy number states from adjacent genomic regions and a HMRF to model correlations between the cancer clones assigned to neighboring spatial locations. c , CalicoST infers allele-specific integer copy numbers for one or more cancer clones, a phylogeny relating these clones, a clone label, an optional tumor proportion for each spot and a phylogeographic model of the spatial expansion of cancer clones.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: a , Accuracy of allele-specific copy numbers across 12 patients from HTAN (WashU cohort) inferred by CalicoST. Each bar represents an inferred cancer clone. b , Length distribution of CNAs identified by CalicoST from SRT data and identified by HATCHet2 from WES for the 9 patients with matched WES data of sufficient tumor purity. Blue bars are CalicoST, and orange bars are HATCHet2, with gray bars indicating the overlap of the two histograms. The median length is 77.4 Mb for CalicoST and 30 Mb for HATCHet2 (vertical dashed lines). c , Allele-specific integer copy numbers inferred by CalicoST from SRT data from a patient with CRC liver metastasis (HT230C1). Rows are cancer clones, and columns are genomic bins. Colors indicate allele-specific copy numbers. d , Allele-specific integer copy numbers inferred by CalicoST from SRT data from a patient with CRC liver metastasis (HT260C1). e , Observed RDR and BAF for chr8 of HT260C1. Points are colored by the inferred allele-specific copy numbers. Horizontal black lines indicate the RDR and BAF of the corresponding copy number states estimated by the HMM. f , Allele-specific integer copy numbers inferred by HATCHet2 from WES data of patient HT260C1. g , RDR and BAF values from WES data for bins from chromosome 8q and bins from other genomic regions with a value of {3,0} copy number state. Black points are expected RDR and BAF values for {3,0} and {2,1} states from HATCHet2 analysis.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: ( a ) CalicoST-inferred cancer clones in PDAC patient HT270P1. Grayscale indicates the inferred tumor proportion within each spot, where more gray indicates a higher proportion of normal cells (lower tumor proportion). ( b ) RDR and BAF along the genome for each inferred clone in HT270P1. Each point represents a genomic bin and is colored by CalicoST-inferred allele-specific copy numbers. The red box highlights a unique deletion in clone 2. ( c – d ) Corresponding plots for PDAC patient HT288P1. Red boxes highlight deletions that are unique to one of the inferred clones.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: H&E images (top) and CalicoST-inferred tumor proportions (bottom) for breast cancer samples: ( a ) HT206B1, ( b ) HT339B1, ( c ) HT268B1, ( d ) HT265B1. The x- and y-axes represent spatial coordinates. The color bar indicates the inferred tumor proportions.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques:

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: The plots for CalicoST include the allele-specific CNAs from all inferred cancer clones, labeled as ‘clone 1’, ‘clone 2’, etc. The plots for HATCHet2, labeled as ‘WES’, are included for the nine patients for whom matched WES data is available and has sufficient tumor purity.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay, Labeling

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: a , b , Accuracy ( a ) and spatial coherence ( b ) comparison among CalicoST, Numbat, InferCNV and STARCH on CRC liver metastasis patient samples. Solid bars indicate predictions of allele-specific copy number states, and dotted bars indicate predictions of total copy number states. c , H&E image of a CRC liver metastasis sample HT260C1. d , Cancer clones inferred by CalicoST. x and y axes are spatial coordinates, and the grayscale represents the proportion of normal cells within each spot, as inferred by RCTD. Other colors indicate cancer clones. e , Cancer clones inferred by Numbat using the same color scheme as in d .

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Comparison, Starch, Clone Assay

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: ( a ) Accuracy of the allele-specific copy number states inferred by CalicoST and Numbat on nine HTAN patients where ‘ground truth’ CNAs were inferred from matched WES data. ( b ) Spatial coherence of the cancer clones inferred by CalicoST and Numbat. The spatial coherence is evaluated by the z-score of joincount statistics, with higher values indicating a greater degree of spatial coherence. Each point represents a cancer clone within each slice of each patient (x-axis). As the two methods identify different numbers of clones, the two boxes include varying numbers of points for each patient. From left to right, the numbers of points in the boxplots are: HT112C1 (6 for CalicoST and 11 for Numbat), HT260C1 (3 for CalicoST and 6 for Numbat), HT265B1 (3 for CalicoST and 4 for Numbat), HT268B1 (10 for CalicoST and 39 for Numbat), HT270P1 (2 for CalicoST and 12 for Numbat), HT288P1 (2 for CalicoST and 6 for Numbat), HT306P1 (2 for CalicoST and 5 for Numbat). The upper and lower bounds of the box denote the 25% and 75% quantiles, the center line denotes the median, and the lower (upper) whiskers denote the smallest (largest) value within 1.5 times the IQR (interquartile range).

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: a , Spatial distribution and phylogeographic tree of three cancer clones inferred by CalicoST in two adjacent slices from patient HT112C1 with CRC liver metastasis. The grayscale indicates the inferred proportion of normal cells within each spot. Diamonds are the spatial centroid of each clone or inferred ancestor, and arrows indicate the inferred directions of tumor development. The distance between two slices in the z -coordinate is enlarged for clearer visualization. b , Allele-specific copy number profiles for the three cancer clones and the corresponding phylogeny (right) with branches in the phylogeny labeled by the number of unique large LOH events that occur on the branch. c , Spatial distribution and phylogeographic tree of two cancer clones inferred by CalicoST in five adjacent slices from patient HT268C1 with breast cancer. Color scheme is the same as a . d , Inferred allele-specific copy numbers and tumor phylogeny.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay, Labeling

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: a , Spatial distribution of cancer clones inferred jointly by CalicoST across five slices from a cancerous prostate. Positioning of five slices is according to ref. . Colors indicate inferred clones, including the normal clone in gray. Arrows represent the phylogeography of tumor evolution. b , Allele-specific copy number profiles for the five cancer clones and the corresponding phylogeny with branches in the phylogeny labeled by the number of unique large LOH events that occur on the branch. Colors indicate allele-specific copy numbers. The orientation and position of triangles indicate mirrored CNA events. c , BAF of each clone in chr6 and chr8. Colors indicate allele-specific copy numbers using the same color scheme as in b .

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay, Labeling

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: ( a ) UMAP of gene expression in spots from five slices of a multi-section prostate cancer patient (without applying any batch effect correction or integration tools). Each point represents a spot, colored by the slice location. ( b ) UMAP of gene expression from five slices, with each spot (point) colored according to the clone assignment from CalicoST. Grayscale indicates the inferred tumor proportion, with more gray representing a higher proportion of normal cells. ( c ) BAF along the genome for spots assigned to clone 5 from three slices (H1 4, H1 5, and H2 5) from the right portion of the prostate. Each point represents a genomic bin, colored by the inferred allele-specific copy numbers from CalicoST.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Gene Expression

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: Each spot is colored by the clone inferred by CalicoST, with gray indicating normal spots. Spots containing the variant allele of the somatic SNV are marked by a black cross. Spots containing the reference allele are marked by a gray circle. The first five SNVs are inferred to be truncal SNVs present in both the left and right sides of the prostate, while the sixth SNV (bottom right) is inferred to be present in only the left side.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Variant Assay

Journal: Nature Methods

Article Title: Inferring allele-specific copy number aberrations and tumor phylogeography from spatially resolved transcriptomics

doi: 10.1038/s41592-024-02438-9

Figure Lengend Snippet: ( a ) Spatial organization of normal (clone 0) and three tumor clones (clones 1-3) inferred by CalicoST on a human melanoma sample sequenced using Slide-tags. ( b ) Compar- ison of cell type labels for each location from and clone labels inferred by CalicoST. ( c ) Allele-specific copy numbers inferred by CalicoST for each clone. ( d ) RDR and BAF along the genome for clones 1 and 2. Colors indicate the allele-specific copy number of the corresponding genomic bin. Red box highlights a LOH event on chr3q that is unique to clone 2.

Article Snippet: We applied CalicoST to infer allele-specific copy numbers on 10x Genomics Visium Spatial Transcriptomics data from 12 patients (26 slices) in HTAN (WashU cohort) across three cancer types (‘Running CalicoST on SRT data’).

Techniques: Clone Assay

Journal: Genome Biology

Article Title: High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper

doi: 10.1186/s13059-025-03773-6

Figure Lengend Snippet: Illustration of STAMapper and its applications. a STAMapper can annotate scST data obtained from mainstream technologies such as image-based and seq-based by leveraging well-annotated sc/snRNA-seq data sequenced from microfluidics-based or droplet-based technologies. b STAMapper models genes and cells as two types of heterogeneous nodes and connects sc/scRNA-seq and scST data by their expression on the shared genes. c STAMapper takes the expression and the heterogeneous relationships of nodes as input. STAMapper then learns embeddings for cells and genes based on the information propagation mechanism on the heterogeneous graph network to fit cell labels from scRNA-seq data by using a graph attention classifier, ultimately utilizing the learned weights of information propagation on the graph to transfer cell labels on spatial data. d The output of STAMapper can be applied for annotation on large-scale scST data, reannotation on scST data, unknown cell-types detection, and gene module extraction

Article Snippet: Fig. 3 Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference.

Techniques: Expressing, Extraction

Journal: Genome Biology

Article Title: High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper

doi: 10.1186/s13059-025-03773-6

Figure Lengend Snippet: Benchmarking cell annotation performance of STAMapper. a Overview of all datasets used for evaluating the performance of STAMapper. We collected 81 single-cell spatial transcriptomics datasets comprising a total of 344 slices, where each dataset is matched with corresponding single-cell transcriptomics data (or scRNA-seq data). b Performance comparison of STAMapper and scANVI, RCTD, Tangram regarding cell annotation accuracy on 81 pairs of scRNA-seq and single-cell spatial transcriptomics datasets. P values were calculated by paired t test. c Performance comparison of the classification accuracies of STAMapper and three other methods on different down-sampling rates (1.0, 0.8, 0.6, 0.4, 0.2) for read counts, where the down-sampling rate of 1.0 means the raw data. The upper panel depicts spatial transcriptomics datasets with more than 200 genes for sequencing (47 datasets), while the lower panel corresponds to fewer than 200 genes (34 datasets)

Article Snippet: Fig. 3 Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference.

Techniques: Single-cell Transcriptomics, Comparison, Sampling, Sequencing

Journal: Genome Biology

Article Title: High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper

doi: 10.1186/s13059-025-03773-6

Figure Lengend Snippet: Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference. AC amacrine cells, EC endothelial cells, MG Müller Glia, PC pericytes, RET reticulocyte, HC retinal horizontal cells, BC bipolar cells, Cones cone cells, RGC retinal ganglion cells, RPE retinal pigment epithelium, Rods Rod cells. c The heatmap of the marker expression for major cell types on the scRNA-seq dataset grouped by manual annotation and on the corresponding spatial transcriptomics dataset annotated by STAMapper, scANVI, and RCTD, respectively. d A schematic illustration of the distribution of cell types within the retina. e. Spatial organization of a slice from spatial transcriptomics dataset corresponding to ( b ), where cells are colored by the annotation by STAMapper, scANVI, and RCTD, respectively

Article Snippet: Fig. 3 Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference.

Techniques: Comparison, Marker, Expressing

Journal: Genome Biology

Article Title: High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper

doi: 10.1186/s13059-025-03773-6

Figure Lengend Snippet: Application of STAMapper to MERFISH hypothalamic dataset. a UMAP plots of mouse hypothalamic dataset colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD, respectively. b Spatial organization of a slice from mouse hypothalamic dataset corresponding to ( a ), cells are colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD, respectively. c Sankey plot showing the accuracy of the cell-type annotation by STAMapper, scANVI, and RCTD, respectively. The left side of the Sankey plot represents manual annotations, while the right side shows the predicted results. The height of each linkage line reflects the number of cells. d Heatmap plot of marker expression for major cell types presented in manual annotation with mismatched labels between manual annotation and STAMapper. e Expression levels of Sema4d (a marker of OD Newly formed ) across different cell types (annotated by STAMapper). f The predicted probability of STAMapper for cells from spatial data (left panel) and unknown cells from spatial data (right panel), the red dash line indicates x = 0.738 in both panels. g Cell-type level distance from spatial data to single-cell data on cell embeddings learned by STAMapper. Bold indicates unknown cells were predicted as this specific cell type, and red denotes cell types present in single-cell data but not annotated in spatial data by manual annotation. h UMAP plots of the co-embedding of scRNA-seq and spatial data learned by STAMapper. Cells are colored by manual annotation, STAMapper prediction without unknown detection, and STAMapper prediction with unknown detection. The percentages in parentheses represent the predicted accuracy

Article Snippet: Fig. 3 Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference.

Techniques: Marker, Expressing

Journal: Genome Biology

Article Title: High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper

doi: 10.1186/s13059-025-03773-6

Figure Lengend Snippet: Application of STAMapper to Nanostring HCC dataset. a UMAP plots of the human HCC dataset colored by STAMapper, scANVI, and RCTD, respectively. Macro macrophage, NK natural killer, DC dendritic cell. b Accuracy of STAMapper, scANVI, RCTD, and Tangram on human HCC dataset. c UMAP plot of the normalized marker expression corresponding to major cell types. d Spatial organization of ROI 1, cells are colored by the annotation of STAMapper, RCTD, and scANVI, respectively. e The normalized marker expression on ROI 1. f Density plot for the distribution of macro cells, annotated by STAMapper. g Physical distance of immune cells to malignant cells, with cells being annotated by STAMapper. h Boxplot for the scores of selected pathways (stemness, proliferation, and MHC-I) on malignant cells near macro and other malignant cells. i Heatmap displaying genes with the highest normalized attention weights, categorized by each cell type. We aggregate the normalized attention weights from that gene to all cells belonging to the cell type and compute their average as the cell type’s normalized attention weights. These scores reflect the gene’s overall contribution to the annotation of that cell type. j Cosine similarity of the gene embedding pairs learned by STAMapper, where gene pairs were TFs collected from hTFtarget. RELA - EPCAM and STAT3 - SMAD3 were validated to exist in HCC malignant cells in the literature

Article Snippet: Fig. 3 Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference.

Techniques: Marker, Expressing

Journal: Genome Biology

Article Title: High-precision cell-type mapping and annotation of single-cell spatial transcriptomics with STAMapper

doi: 10.1186/s13059-025-03773-6

Figure Lengend Snippet: Application of STAMapper to Slide-tags human prefrontal cortex dataset. a , b UMAP plot for the co-embedding of scRNA-seq and spatial dataset learned by STAMapper. Cells are colored based on the prediction of STAMapper ( a ) and the origin of the datasets ( b ), respectively. Oligo oligodendrocytes, OPC oligodendrocyte progenitor cells. c The predicted cell-type probabilities for each cell (each column) in the spatial data. A maximum of 50 cells was subsampled from each type for visualization. d UMAP plots showing the co-embedding of the scRNA-seq and spatial dataset learned by STAMapper, cells are colored by the normalized expression levels of GPR17 . e Boxplots of the cosine similarity between gene embedding pairs grouped by the number of shared pathways. f UMAP plot for the distribution of gene embedding. Genes are colored by clusters identified through the Leiden algorithm. g Abstracted graph of the heterogenous cell-gene graph, where nodes represent cell types (pink) or gene modules (blue). Node size reflects the number of cells in a cell type or genes in a module. Edge width varies with the average expression levels of cell types linked to gene modules, determined by STAMapper. h UMAP plots showing the co-embedding of scRNA-seq and spatial dataset learned by STAMapper, cells are colored by the normalized expression levels of Module 12. i Enrichment analysis of gene module 12 related to Oligo_GPR17 cells. j Spatial organization of cells from the spatial dataset. Cells are clustered by STAGATE with resolution = 0.05. k The Normalized expression of SYT4 (marker gene of grey matter) and LRP2 (marker gene of white matter). l – n Spatial organization and UMAP plot of astrocyte ( l ), excitatory ( m ), and inhibitory ( n ) Subtypes predicted by STAMapper from the spatial dataset, Subtypes with more than 20 cells are shown. The UMAP coordinates are calculated from the expression of the spatial data

Article Snippet: Fig. 3 Application of STAMapper to MERFISH retina datasets. a Performance comparison of STAMapper and scANVI, RCTD, Tangram, where each box represents the method’s performance on the 50 paired datasets (five scRNA-seq datasets and ten single-cell spatial transcriptomics datasets). b UMAP plots of mouse retinal dataset (VZG105a_WT3) cells colored by the manual annotation and the prediction of STAMapper, scANVI, and RCTD using the mouse_LD_60 scRNA-seq dataset as the reference.

Techniques: Expressing, Marker

Journal: Journal of Translational Medicine

Article Title: Transformative advances in single-cell omics: a comprehensive review of foundation models, multimodal integration and computational ecosystems

doi: 10.1186/s12967-025-07091-0

Figure Lengend Snippet: Overview of Methodological Workflows for Multi-Omics and Spatial Transcriptomics Analysis. a Nicheformer Model for Gene Expression Integration: The Nicheformer model processes tokenized gene expression data and assay-specific markers using transformer embeddings, producing unified outputs for gene ranking and modality integration. This enables accurate predictions for gene regulatory networks (GRN) and drug response analysis . b LocalCLiP for Spatial Transcriptomics: LocalCLiP utilizes a local transformer model to integrate spatial transcriptomics data, using KNN for image patch analysis and gene expression prediction, providing insights into tissue-specific molecular patterns . c BioTask Executor for Task-Specific Analysis: The BioTask Executor handles various biological tasks, from zero-shot learning to GRN inference and drug response prediction, by preprocessing data, initializing pretrained models (e.g., SCGPT, Geneformer), and fine-tuning them for task-specific applications . d Human-8CATAC-CorpuS for Multi-Tissue Analysis: The Human-8CATAC-CorpuS dataset, with 5 million cells from 31 tissues, is used to train models for gene expression prediction and cCRE signal reconstruction, enabling comprehensive analysis of tissue-specific regulatory elements . The schematics were adapted from [ , , ] and

Article Snippet: This enables accurate predictions for gene regulatory networks (GRN) and drug response analysis [ ]. b LocalCLiP for Spatial Transcriptomics: LocalCLiP utilizes a local transformer model to integrate spatial transcriptomics data, using KNN for image patch analysis and gene expression prediction, providing insights into tissue-specific molecular patterns [ ]. c BioTask Executor for Task-Specific Analysis: The BioTask Executor handles various biological tasks, from zero-shot learning to GRN inference and drug response prediction, by preprocessing data, initializing pretrained models (e.g., SCGPT, Geneformer), and fine-tuning them for task-specific applications [ ]. d Human-8CATAC-CorpuS for Multi-Tissue Analysis: The Human-8CATAC-CorpuS dataset, with 5 million cells from 31 tissues, is used to train models for gene expression prediction and cCRE signal reconstruction, enabling comprehensive analysis of tissue-specific regulatory elements [ ].

Techniques: Biomarker Discovery, Gene Expression