Journal: bioRxiv

Article Title: Spartan: Spatial Activation Aware Transcriptomic Analysis Network

doi: 10.64898/2026.02.18.706570

Figure Lengend Snippet: (A–B) Spatial domain identification for MERFISH 0.19 (SDMBench) and Vizgen MERFISH S2R3 samples. Ground-truth annotations (left) are compared with clustering results from Spartan (v1, v2), NichePCA, BANKSY, and SpatialLeiden (v1, v2). ARI values above each map quantify agreement with ground truth. (C) Spatially variable gene (SVG) examples for the MERFISH 0.19 sample. Spatial expression patterns align with anatomical boundaries and support the cluster assignments in (A). (D) Stereo-seq mouse embryo (E9.5E2S2). Ground-truth annotations (left) and spatial expression of representative embryo marker genes (right) show anatomically consistent localization. (E) SVGs for the Vizgen MERFISH mouse brain S2R3 sample, highlighting region- and cell-type–specific expression patterns. (F) 10x Visium human dorsolateral prefrontal cortex (DLPFC) sample 151673. Ground-truth cortical layers (left) and layer-enriched marker genes (right) demonstrate accurate laminar recovery.

Article Snippet: The Vizgen MERFISH mouse brain dataset further highlights Spartan’s ability to capture regionally specialized neuronal and glial populations across the adult brain ( ; Supplementary Fig. 3C).

Techniques: Expressing, Marker

Journal: bioRxiv

Article Title: Spartan: Spatial Activation Aware Transcriptomic Analysis Network

doi: 10.64898/2026.02.18.706570

Figure Lengend Snippet: (A–B) Spatial domain identification for MERFISH 0.19 (SDMBench) and Vizgen MERFISH S2R3 samples. Ground-truth annotations (left) are compared with clustering results from Spartan (v1, v2), NichePCA, BANKSY, and SpatialLeiden (v1, v2). ARI values above each map quantify agreement with ground truth. (C) Spatially variable gene (SVG) examples for the MERFISH 0.19 sample. Spatial expression patterns align with anatomical boundaries and support the cluster assignments in (A). (D) Stereo-seq mouse embryo (E9.5E2S2). Ground-truth annotations (left) and spatial expression of representative embryo marker genes (right) show anatomically consistent localization. (E) SVGs for the Vizgen MERFISH mouse brain S2R3 sample, highlighting region- and cell-type–specific expression patterns. (F) 10x Visium human dorsolateral prefrontal cortex (DLPFC) sample 151673. Ground-truth cortical layers (left) and layer-enriched marker genes (right) demonstrate accurate laminar recovery.

Article Snippet: The data were acquired from the following websites or accession numbers: (1) Stereo-seq, BaristaSeq, MER-FISH, osmFISH, STARmap, STARmap*, 10x Visium datasets with regional annotations were downloaded from SDMBench; (2) Vizgen MERFISH dataset was downloaded from Vizgen MERFISH Mouse Brain Receptor Map; (3) The Visium HD dataset generated in this study is available in ArrayExpress under accession number E-MTAB-16671.

Techniques: Expressing, Marker

Journal: bioRxiv

Article Title: SpaceTrooper : a quality control framework for imaging-based spatial omics data

doi: 10.64898/2025.12.24.696336

Figure Lengend Snippet: SpaceTrooper QS is generalizable across spatial transcriptomics platforms. A. Spatial map of the quality score across the Xenium human lung cancer dataset, with highlighted two regions within a peripheral vessel (boxed regions i and iii), one region in an isolated tissue fragment (boxed region ii), and one region within the tissue mass (boxed region iv). B. Spatial distribution of the QS, signal density, and composite fluorescence signals in boxed region i. White arrows indicate technical artifacts as out-of-focus regions, uneven cytoplasmic (interior protein and RNA) and membrane (boundary) staining, and segmentation inaccuracies. C. Same as in B for the area highlighted in boxed region iii. D. Spatial distribution of the quality score in the Vizgen MERFISH mouse liver dataset. Light blue boxes indicate examples of peripheral and interior regions comprising low-QS cells. E. Magnified view of a peripheral low-QS region located at the left side of the image. Reduced QS in this area (top) is primarily driven by diminished signal density (bottom). F. Same as in E for an interior low-QS region. In panels D , E , and F , cells shown in grey correspond to cells with zero counts.

Article Snippet: SpaceTrooper generalizability across imaging-based spatial omics technologies and tissue types was further evaluated using the Vizgen MERFISH mouse liver dataset.

Techniques: Isolation, Fluorescence, Membrane, Staining