Dissecting glial scar formation by spatial point pattern and topological data analysis

These data were generated by the Laboratory of Neurovascular Interactions (https://elalilab.com/) at University Laval (Quebec, Canada), and reported in "Dissecting glial scar formation by spatial point pattern and topological data analysis". 

Please refer to the Open Science Framework (OSF) repository (https://osf.io/3vg8j/) or GitHub (https://github.com/elalilab/GlialScar_PPA-TDA_2022) to see the processing pipeline.

AUTHORS
Manrique-Castano, Daniel; Bhaskar, Dhananjay; ElAli, Ayman

KEYWORDS
Stroke, cerebral ischemia, brain injury, glial scar, reactive astrocytes, reactive microglia, 

1. STUDY DESCRIPTION                                                             
This research provides a quantitative analysis of reactive glia and glial scar formation in a mouse model of cerebral ischemia. The dataset in this repository consists of raw widefield microscopy images from healthy and ischemic animals.    

2. EXPERIMENTAL CONDITIONS
Six-month-old C57BL/6 mice were subjected to 30 minutes of cerebral ischemia by middle cerebral artery occlusion (MCAO). Brains were harvested at 5, 15, and 30 days post-ischemia (DPI) (see 10.5281/zenodo.3559570). 5 sham animals were included as controls. The full protocol for brain harvesting is available at 10.17504/protocols.io.4r3l27q5pg1y/v1. Brain sections were stained with NeuN, Gfap, and Iba1 antibodies to detect neurons and reactive glia after injury. Full protocol available at 10.17504/protocols.io.yxmvmk94og3p/v1 
                        
3. FILE DESCRIPTION

- GT5X_Gfap_Iba1_NeuN.rar: Contain widefield (5x magnification) .tif images grouped by animals (5-7 images per animal; see research article for further details). The images were taken with the following parameters.

Objective:        Fluar 5x/0.25 M27
Scaling per pixel:    1.300 x 1.300 µm
Bit depth:        16 bit    

Stainings:
Neun            Channel    AF647; Excitation 653; Emission    668; Exposure 3 s
IBA1            Channel    AFCy3; Excitation 458; Emission    561; Exposure 4 s
GFAP            Channel    AF488; Excitation 493; Emission    517; Exposure 1 s
DAPI            Channel    AF405; Excitation 353; Emission    465; Exposure 50 ms

We used a FIJI script to pre-process the original .czi files. The script is shared in the GitHub repository under the name GT_Exp2_5x_GenerateTiffs.jim.

- GT10X_Gfap_Iba1_NeuN.rar: Contain a single widefield (10x magnification) .tif image per animal at the level of the MCA territory (see research article for further details). The images were taken with the following parameters.

Objective:        ECM paln-NeoFluar 10x/0.30 M27
Scaling per pixel:    0.45 x 0.45 µm
Bit depth:        16 bit    

Stainings:
Neun            Channel    AF647; Excitation 653; Emission    668; Exposure 200 ms
IBA1            Channel    AFCy3; Excitation 458; Emission    561; Exposure 250 ms
GFAP            Channel    AF488; Excitation 493; Emission    517; Exposure 100 ms
DAPI            Channel    AF405; Excitation 353; Emission    465; Exposure 10 ms

We used a FIJI script to pre-process the original .czi files. The script is shared in the GitHub repository under the name GT_Exp2_10x_GenerateTiffs.jim.
    
For 5x and 10x images, the following naming strings apply:

GT5x:         Research project identifier indicating the magnification
M01(n):     Animal ID
5D(n):      Days post-ischemia. 0D refers to healthy (naive) animals. 
Scene1(n):     Bregma level. Scene 1 corresponds to the most anterior area sampled, while Scene 6 or 7 is the most posterior.

- PointPatterns_10x.rds: 2D point patterns of GFAP, IBA1, and NeuN generated by the r-package spatstat. The observation window comprises a horizontal ROI from the ventricular area to the outer border of the dorsolateral cerebral cortex. The point patterns were generated from the files and coordinates contained in the QupathProjects_10x.rar file in this repository. To reproduce the generation of point patterns please refer to the associated GitHub repository (https://github.com/elalilab/Stroke_GlialScar_PPA-TDA). 

- PointPatterns_5x.rds: 2D point patterns of GFAP, IBA1, and NeuN generated by the r-package spatstat. The observation window comprises the ischemic hemisphere. The point patterns were generated from the files and coordinates contained in the QupathProjects_5x.rar file in this repository. To reproduce the generation of point patterns please refer to the associated GitHub repository (https://github.com/elalilab/Stroke_GlialScar_PPA-TDA). 

- QupathProjects_5x.rar: QuPath project folder for 5x images (GT5X_Gfap_Iba1_NeuN.rar). Each subfolder (per animal) contains the necessary files to import annotations (alignment to the Allen Brain Atlas) generated by ABBA (https://biop.github.io/ijp-imagetoatlas/). Please see the research article for further details. 

**NOTE** Gfap, Iba1, and NeuN folders contain raw .tsv data originated by QuPath (cell counting). These folders are read in the R processing pipeline to extract the coordinates of each cell. Please make sure the whole folder is in the R working directory. The file "project.qpproj" in each folder opens the QuPath project in QuPath and reads the classifiers and data folders. Each folder also contains "_Alignement.json" and "_Registration_json" files generated during the alignment and annotation procedures in ABBA. However, when the route of the source images is changed, the plugin does not allow rerouting, and the files are of no practical use. The issue has been reported to the ABBA Github repository.   

- QupathProjects_10x.rar: QuPath project folder for 10x images (GT5X_Gfap_Iba1_NeuN.rar). The folder contains the necessary files to import annotations (Alignment to the Allen Brain Atlas) generated by ABBA (https://biop.github.io/ijp-imagetoatlas/). Please see the research article for further details. 

**NOTE** Gfap, Iba1, NeuN, and DAPI folders contain raw .tsv data originated by QuPath (cell counting). These folders are read in the R processing pipeline to extract the coordinates of each cell. Please make sure the whole folder is in the R working directory. The file "project.qpproj" opens the QuPath project in QuPath and reads the classifiers and data folders.