CellColoc example data

Example data for CellColoc This Zenodo record contains example and validation data for testing, demonstrating, and reproducing analyses performed with CellColoc, an open-source, script-driven workflow for image-based cell segmentation, object-based marker colocalization, regional occupancy analysis, and per-cell feature extraction. The record contains two dataset archives: A small example-data archive for quick testing and tutorials. An extended preprint-validation archive containing the full biological microglia example dataset and two synthetic benchmark datasets with ground truth. Dataset contents Small example-data archive The small archive is intended for lightweight testing of CellColoc installation, image loading, channel assignment, marker handling, and basic colocalization workflows. microglia_3D/: reduced multichannel CZI microglia example dataset for quick CellColoc testing. dapi_stained_nuclei_2D/: redistributed 2D fluorescence microscopy example data from Rathar (2018), included for lightweight testing and demonstration. Extended preprint-validation archive The extended archive contains the datasets used to reproduce the analyses and figure panels reported in the CellColoc preprint. microglia_3D_full/: extended biological microglia CZI dataset used for the preprint biological example. synthetic_benchmark_data_sharp/: sharp filled-object synthetic benchmark dataset with known ground truth. synthetic_benchmark_data/: Gaussian synthetic benchmark variant with known ground truth. The extended archive also includes CellColoc result tables, masks, analysis configuration files, and manifests generated by the preprint analysis scripts. These files allow the numerical results and preprint figure panels to be inspected without rerunning the full segmentation workflows. Biological microglia dataset The microglia_3D_full/ dataset contains Zeiss CZI confocal image stacks from adult C57BL/6J mice carrying the reporter genotype: Cx3cr1-creER cre/wt ; Rosa26_tdTom fl/wt The dataset contains hippocampal CA1 and cortical imaging fields. Animals received either tamoxifen or vehicle treatment before tissue collection. The treatment annotation is included to support the preprint sensitivity analysis comparing vehicle- and tamoxifen-treated animals. Channel Marker Description 0 Cx3cr1-tdTomato Genetically labelled CX3CR1-positive cells; in brain parenchyma, these are expected to be predominantly microglia. 1 Iba1 immunostaining Independent microglia / macrophage-lineage marker used to assess whether tdTomato-positive cells are Iba1-positive. 2 DAPI Nuclear stain used for anatomical orientation and tissue-quality assessment. The main CellColoc analysis compares tdTomato-positive cells in channel 0 against Iba1 signal in channel 1. The DAPI channel is included as supporting anatomical information. Representative imaging metadata report Zeiss laser-scanning confocal microscopy, a Plan-Apochromat 20x/0.8 M27 air objective, three channels, 1024 x 1024 pixels, commonly 30 z-slices, and physical pixel sizes of approximately 0.624 micrometers in x/y and 1.0 micrometer in z. Exact metadata should be read from the individual CZI files. Synthetic benchmark datasets The synthetic benchmark datasets contain 20 three-channel OME-TIFF stacks each. They were generated to provide controlled validation data with known ground truth for object counts, object-level morphology, marker positivity, and regional occupancy. synthetic_benchmark_data_sharp/ This is the main synthetic benchmark dataset used in the CellColoc preprint. Channels 0 and 1 contain sharp filled cell-like objects with explicit object boundaries. Channel 1 objects were generated partly from channel 0 object locations so that approximately 70% of channel 0 cells have a marker-channel overlap and approximately 30% do not. Channel 2 contains an irregular non-cellular region occupying approximately 30% of the image area. The Cellpose-based result variant in cellpose_variant_stacks/ was used for the main synthetic benchmark figure in the CellColoc preprint. synthetic_benchmark_data/ This is the Gaussian synthetic benchmark variant. Channels 0 and 1 contain softer Gaussian-blurred cell-like objects with noise. It is included as a supplementary validation dataset to evaluate CellColoc behavior when object boundaries are less sharply defined. Ground truth Both synthetic benchmark folders include ground-truth tables in ground_truth/: synthetic_benchmark_ground_truth_summary.csv synthetic_benchmark_ground_truth_objects.csv synthetic_benchmark_ground_truth_colocalization.csv labels/ containing ground-truth label masks The included CellColoc results can be used to reproduce benchmark plots and compare CellColoc outputs against known ground truth. Preprint reproducibility The extended datasets include outputs generated by the CellColoc preprint scripts, including segmentation masks, object tables, ROI summaries, analysis configuration files, and batch manifests. additional_scripts/create_synthetic_benchmark_data.py additional_scripts/run_synthetic_benchmark_cellcoloc.py additional_scripts/run_synthetic_benchmark_cellcoloc_cellpose.py additional_scripts/run_microglia_3d_full_batch.py additional_scripts/preprint_generate_figures.py These files are included in CellColoc's official GitHub repository to support transparent inspection and reproduction of the CellColoc preprint analyses. Redistributed Zenodo example data from Rathar (2018) This record also includes example image data from: Rathar, Raissa. “DAPI stained nuclei more or less clustered.” Zenodo, Version 1.0.0, 2018.DOI: 10.5281/zenodo.1304211License: Creative Commons Attribution 4.0 International, CC BY 4.0. The Rathar dataset contains fluorescence microscopy images of DAPI-stained nuclei and corresponding TexasRed-stained cells. These files are redistributed here as example material for CellColoc testing and demonstration. Modifications to the Rathar dataset files, if any, are limited to file selection, file organization, and integration into the CellColoc example-data structure. The biological image content of the original Rathar files was not intentionally altered. Attribution and license information Files originating from Rathar (2018) are redistributed under the Creative Commons Attribution 4.0 International license. Please cite the original dataset when using these files: Rathar, Raissa. “DAPI stained nuclei more or less clustered.” Zenodo, 2018. DOI: 10.5281/zenodo.1304211. The remaining files were provided as CellColoc example and validation data by the authors of CellColoc. Please cite this Zenodo record and the associated CellColoc software/preprint when using these files. Intended use These data are intended for: testing CellColoc installation and input handling; demonstrating multichannel fluorescence image loading; testing marker-channel assignment; demonstrating object-based cell-marker colocalization workflows; validating synthetic benchmark analyses against known ground truth; reproducing the CellColoc preprint example analyses and figure panels; inspecting saved masks, per-cell tables, per-ROI summaries, and analysis configuration files. These files are provided as example and validation data for CellColoc. They should not be interpreted as a complete standalone biological study.