Kim2025_VisiumMicrogliaMorphology

Created: 16 September, 2025 Last updated: 24 September, 2025

This github repo contains the ImageJ and R scripts used for the integration of publicly available Visium-SPG data and morphology analysis presented in Kim et al., 2025:

insert citation

Below are the steps that were taken for this analysis, and links & locations of where you can find related scripts (Github) and datasets (OSF):

Download publicly available datasets from Huuki Myers 2024

• spe object – Github
• Visium-SPG .tiff files
• Visium-SPG spot coordinates
• Visium spot deconvolution data → tidied and saved as .rds object: – OSF: rds_files/SpotDeconvolution_A1-C1.rds

Prepare images for MicrogliaMorphology

• use ImageJ to manually subtract lipofuscin.tiff from tmem119.tiff – OSF: MicrogliaMorphology_InputImages_LipofuscinSubtracted
  • open paired lipofuscin and tmem119 channel images
  • Process > Image Calculator: lipofuscin.tiff - dapi.tiff

Run MicrogliaMorphology and MicrogliaMorphologyR

• run updated thresholding macro – Github
• run rest of MicrogliaMorphology steps
• explore data using MicrogliaMorphologyR – Github

Save morphology data

• save dataframe of microglia and morphology features – Github, OSF: rds_files/MicrogliaMorphology_FinalData.rds
• save .csv files of CellIDs for each image

Save ROI .zip files of microglia ROIs and Visium Spot ROIs using following ImageJ macros

• SavingJustCellROIs.ijm – Github
• VisiumSpotOverlay.ijm - Github, OSF: ImageJ_ROI_ZipFiles/A1-D1_VisiumSpot_ROIs

Generate microglia information sheet, which includes new ImageJ measures: proportion_cellarea, proportion_spotarea

• AreaOverlap.ijm: ImageJ macro to calculate area overlap between visium spots and microglia cells - Github
• manually calculate cell areas using microglia masks
  • open lipofuscin-subtracted .tiff image - OSF: MicrogliaMorphology_InputImages_LipofuscinSubtracted
  • drag and drop image-specific microglia ROI .zip file - OSF: ImageJ_ROI_ZipFiles/A1-C1_AllMicroglia_PreFiltering_ROIs.zip
  • select all in ROI manager and click Measure button
    • make sure Analyze > Set Measurements has ‘Area’ and ‘Display Label’ selected
  • save cell area measurement output for each image separately
• read in AreaOverlap.ijm output, cell areas, and calculate proportion_cellarea and proportion_spotarea measures - Github
• save to Microglia-spot-info dataframe to be read in later for integration with Visium-SPG object – OSF: rds_files/Microglia-spot-info_LipofuscinSubtracted.rds

Analysis of antibody fluorescent channel signals at Visium capture area level (Fig. S1A)

• manually calculate mean signal intensity histogram for each antibody for each image
  • open lipofuscin-subtracted .tiff image - OSF: MicrogliaMorphology_InputImages_LipofuscinSubtracted
  • drag and drop image-specific Visium capture area ROI file - OSF: ImageJ_ROI_ZipFiles/A1-D1_VisiumCaptureAreas.zip
  • select ROI in ROI manager
  • select Analyze > Histogram, then click List button to obtain histogram data
  • save histogram data output for each image individually
• files joined and saved as VisiumCaptureAreas_MeanGrayValues_AllChannels.rds – Github, OSF: rds_files/VisiumCaptureAreas_MeanGrayValues_AllChannels.rds
• signal intensity histograms plotted for capture area-specific antibody signals – Github

Analysis of antibody fluorescent channel signals at Visium Spot level (Fig. S1B-D)

• manually calculate mean signal intensity and area for each antibody for each image
  • open lipofuscin-subtracted .tiff image - OSF: MicrogliaMorphology_InputImages_LipofuscinSubtracted
  • drag and drop image-specific Visium spot ROIs - OSF: ImageJ_ROI_ZipFiles/A1-D1_VisiumSpot_ROIs
  • select all in ROI manager and click Measure button
    • make sure Analyze > Set Measurements has ‘Area fraction’, ‘Mean gray value’, and ‘Display Label’ selected
  • save mean signal intensity and area measurements as separate file for each image
• files read in and saved as MeanGrayValue_AllChannels.rds – Github, OSF: rds_files/MeanGrayValues_AllChannels.rds
• signal intensity and percentage area histograms plotted for spot-level antibody signals – Github
• pearson’s correlations calculated across mean signal intensity or signal area with cell type proportions/numbers in spe object – Github

Identify multi-cell spots (excluded 8 ‘problem’ masked spots)

• R code to identify any spots with multiple morphology classes that were masked/hidden in filtering process (Fig. 2A; Fig. S3) - Github, OSF: rds_files/MulticlassSpotIDs.csv, AllMGdata_preANYfiltering.csv, MultiClassSpots_ExcludeAtEnd.csv, PseudobulkedSpots.csv

PanglaoDB marker gene analysis - antibody features correlations (Fig. S2A)

• download and save .rds files of PanglaoDB marker gene lists - Github, OSF: rds_files/PanglaoDB_markergenes_celltype
• plot correlations of marker genes with signal intensity and percentage area at spot level - Github

Area overlap analysis to generate new measures: proportion_spotarea and proportion_cellarea

• AreaOverlap.ijm – Github
• R code to create Microglia-spot-info file (mgdata in scripts) - Github

Pseudobulk exploration (Figs. 2D-E, S2B-C)

• generate pseudobulked data – ADD GITHUB, OSF: rds_files/Pseudobulk_final.rds
• WM vs. GM proportions (Fig. 2D-E) – Github
• WM vs. GM gene overlap enrichment against PanglaoDB lists - Github

NMF, DE analysis

• Github

MGEnrichment, ErmineJ enrichment

• Github

CD74 and SPP1 analysis

• Github