Allen Brain Atlas Counts Database:							
S.Manju Attili

The ABA Counts database contains all the necessary files and scripts needed by researchers for further data mining/analysis. It is organized into 4 folders as follows:

1. Masked_Segmentations: Folder contains 41 subdirectories of masked segmented images generated for each of the 654 masked images from the Allen’s reference atlas. The images are organized in subdirectories named after their corresponding parcels.

2. Output_Files: Folder contains 2 subdirectories, ‘ImageJ_output_files’ and ‘CellProfiler_output_files’. These subdirectories contain output files in the form of excel spreadsheets from each of the two pipelines used (ImageJ and CellProfiler). No corrections have been applied to this data except declumping.

3. Scripts: This folder contains the following scripts:

a. ImageJ_Processing: ImageJ script used to process all the coronal images for acquiring initial cell counts. This script includes the Watershed segmentation algorithm application. Output is in the form of spreadsheets (output_files).

b. CellProfiler_Processing: CellProfiler project file containing calibrations and settings used to obtain initial cell counts from the coronal images. This file includes the settings for segmentation of clumped cells. Output is in the form of spreadsheets (output_files).

c. Abercrombie_Correction: R script for applying the Abercrombie algorithm on the output files. Input for the program is the set of output files generated in the form of spreadsheets.

d. Borders_Correction: R script for applying border correction on the output files.

e. Mask_Areas: ImageJ script for calculating mask areas for the images.

f. Spatial_Random: ImageJ script used to generate random coordinates from the coronal images. Output is in the form of spreadsheets.

g. Spatial_Real: ImageJ script used to generate real coordinates from the coronal images. Output is in the form of spreadsheets.

h. Spatial_Analysis: R script used for spatial distribution analysis. It does three things:

- Calculates nearest neighbor distances for a given file with x and y coordinates and generates new file with the nearest neighbor distance value for each file; input files with real and random coordinates need to be generated (scripts available on the database, ‘spatial_real, spatial_random ImageJ scripts’).  

- Next, it calculates the average nearest neighbor distance and standard deviation and generates a report with these values for each file; input for this part is the set of files generated from part i.

- Finally, it conducts t-tests for each file using the averages and standard deviations from real and random values and generates a report with the corresponding p-values. The input for the part iii is a consolidated spreadsheet that can be generated by combining real and random reports generated from part ii.

Note: All scripts are written for batch processing.

4. Stereoinvestigator_Files: Folder contains 65 excel files (results) and 65 DAT (data) files organized into high level parcels into 30 subdirectories and an additional ‘Stereology_vs_Computational’ spreadsheet that compares and plots counts from the two methods.