Imaging Mass Cytometry Dataset of Human Carotid Atherosclerotic Plaques

Le résumé fourni par la source

Overview This repository contains de-identified raw and processed imaging mass cytometry (IMC) data supporting the study entitled “Spatial Single-Cell Proteomics Identifies Macrophage Niches Driving Human Carotid Plaque Instability.” The repository includes pixel-level raw IMC acquisition data, processed multichannel images, cell-segmentation masks, extracted single-cell measurements, spatial-neighbor information, sample and ROI metadata, and processed single-cell objects. The study was organized into a discovery and cell-annotation cohort (Cohort 1) and an independent validation and refinement cohort (Cohort 2), profiled using three complementary IMC antibody panels. Study design and analytical roles of the cohorts and antibody panels Cohort 1: discovery and cell-annotation cohort Cohort 1 included 28 arterial plaque specimens, comprising 13 carotid and 15 femoral artery samples. Data from this cohort were generated using two complementary IMC antibody panels. Panel 1 was designed for broad phenotyping of major immune, stromal, endothelial, and vascular smooth muscle cell populations and was used to establish the spatial single-cell atlas of human atherosclerotic plaques. Panel 2 provided more detailed characterization of macrophage and other myeloid molecular and functional states and their spatial organization. Samples from both arterial sites were processed for cell segmentation and contributed to cell-subpopulation definition and annotation. However, only carotid plaque samples were included in the downstream analyses and figures concerning carotid plaque organization and instability presented in the manuscript. Femoral samples were used to support cell-subpopulation annotation and were not included in the statistical testing of the reported carotid plaque associations. Both the complete Cohort 1 datasets and their corresponding carotid-only subsets are provided to document the annotation workflow and facilitate reproduction of the reported carotid plaque analyses. Cohort 2: independent validation and refinement cohort Cohort 2 comprised 62 human carotid atherosclerotic plaque specimens obtained from 47 patients, encompassing 227 regions of interest (ROIs) and 543,113 segmented cells. This cohort was profiled using Panel 3, which incorporated markers prioritized from the discovery analyses together with additional markers for higher-resolution macrophage phenotyping. In particular, Panel 3 enabled further distinction between MMP9⁺SPP1⁺ remodeling-associated macrophages and PLIN2⁺SPP1⁻ foam-like macrophages. Cohort 2 was used to refine and validate the cellular states and spatial organizations identified in Cohort 1 and to assess their associations with histopathological features, Carotid-RADS severity, and imaging- and pathology-defined plaque instability. Relationship among the three antibody panels The three antibody panels were designed for complementary analytical purposes and differed in marker composition: Panel Cohort Primary analytical purpose Panel 1 Cohort 1 Broad identification of major immune, stromal, endothelial, and vascular smooth muscle cell populations and construction of the initial spatial cellular atlas Panel 2 Cohort 1 Detailed characterization of macrophage and other myeloid molecular and functional states Panel 3 Cohort 2 Higher-resolution macrophage phenotyping and validation of associations with plaque histopathology, Carotid-RADS, and plaque instability The panels should not be regarded as technical replicates. Marker intensities and panel-specific cell annotations should not be directly pooled or quantitatively compared across panels without an appropriate cross-panel harmonization strategy. Likewise, failure to identify a particular cell state in another panel should not be interpreted as biological absence when the markers required to define that state were not included in the corresponding panel. Repository organization The data are organized into three archives with broadly similar internal structures. The principal differences are the study cohort, antibody panel, and processed single-cell objects included in each archive. Archive Study role and contents 001_Cohort_1_Panel1.zip Cohort 1 discovery data acquired using Panel 1 for broad phenotyping of major immune, stromal, endothelial, and vascular smooth muscle cell populations 002_Cohort_1_Panel2.zip Cohort 1 discovery data acquired using Panel 2 for detailed characterization of macrophage and other myeloid molecular and functional states 003_Cohort_2_Panel3.zip Cohort 2 data acquired using Panel 3 for higher-resolution macrophage phenotyping and independent validation of associations with plaque pathological features and instability Raw data and confidentiality Pixel-level raw IMC signals were generated as TXT files concurrently with the corresponding proprietary MCD files during data acquisition. Each TXT file represents an individual ROI and contains pixel coordinates and raw ion-count intensities for all acquired metal channels. Only de-identified TXT acquisition files are deposited. The corresponding MCD containers were excluded because their embedded acquisition metadata contained potentially identifiable patient information. The deposited TXT files retain the pixel-level channel-intensity information required to reconstruct and reprocess the released ROIs. All deposited sample, specimen, and ROI identifiers are study-specific pseudonyms rather than direct patient identifiers. IMC data processing and analysis Upstream IMC image processing was primarily based on the Bodenmiller laboratory IMC workflow and implemented using the Docker-containerized steinbock toolkit. Image denoising was performed using IMC-Denoise, developed by Lu et al. (Nature Communications, 2023; doi:10.1038/s41467-023-37123-6 ). Cell segmentation was performed using a combined DeepCell–Cellpose2 workflow. The resulting cell masks were used to extract single-cell marker intensities, morphological features, centroid coordinates, and cell–cell spatial-neighbor relationships. CellProfiler was used for relevant cell-level feature extraction steps. Downstream analyses were primarily performed in R and included: Cell-population and cell-subpopulation annotation Characterization of macrophage and other myeloid states Cellular-neighborhood identification Spatial-interaction analysis Quantification of cell composition across ROIs Associations with plaque histopathological features Associations with Carotid-RADS and plaque instability Processed single-cell data were organized as SpatialExperiment objects and saved in .qs format. Shared file and folder descriptions Internal file and folder names vary slightly among the three archives. The following descriptions summarize the corresponding data components. raw_text/ raw / raw_txt.zip:** De-identified pixel-level raw IMC TXT acquisition files. Each file corresponds to an individual ROI and contains spatial coordinates and raw ion-count intensities for all acquired channels. img: Preprocessed multichannel IMC images used for downstream segmentation and image analysis. masks: Final cell-segmentation masks in which individual integer labels identify individual segmented cells. intensities: Single-cell marker-intensity measurements extracted from the corresponding multichannel IMC images and segmentation masks. regionprops: Cell-level morphological and spatial measurements, including cell area, shape-related features, and centroid coordinates. neighbors: Cell–cell spatial-neighbor relationships used for downstream cellular-neighborhood and spatial-interaction analyses. compensation: Files associated with correction of channel spillover during IMC preprocessing. cellprofiler_input/ cellprofiler_output:** Input and output files associated with CellProfiler-based extraction of cell-level measurements. cell_measurement.cppipe: CellProfiler pipeline used for cell-level feature extraction. panel: Antibody-panel and metal-channel metadata associated with the c

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.