Cellpose-Quantification

A repository for single-cell quantification using mask formats generated by Cellpose! This CLI tool enables the analysis of single-cell data by enabling the quantification of signal intensities and cell-specific features from images and corresponding masks.

Getting Started

Installing Dependancies

• Clone this git repository locally
• Setup Python environment and install packages

pip install -r requirements.txt

Executing program

This program can be run from the command line with the following example command:

python RunQuantification-CLI.py ./path/to/images ./path/to/masks markers.csv 

With Normalisation

python RunQuantification-CLI.py ./path/to/images ./path/to/masks markers.csv --norm log

Expected Parameters

• ./path/to/image
  • Can take Input from a directory containing tiff images
• ./path/to/masks
  • Can take a directory .npy, .png, and .tif masks as inputs generated by cellpose
• markers.csv
  • The path to metadata relating to markers corresponding to each markers, formatted on one row e.g. CD45, DAPI, Ki67, etc
• --norm
  • An optional parameter which defines a normalisation technique to use, current implementation allows: minmax and log normalisation.

Expected Output

The output is a CSV file containing quantitative information for each segmented cell. Below is an example:

Cell_ID	Nuclear	CD31	CK5	SMA	Ki67	CK8	CCASP3	area	centroid-0	centroid-1	perimeter	eccentricity	solidity	orientation	Filename
1	12.622871	0.600973	0.571776	0	0.238443	0.002433	0.381995	411	200.211679	697.878345	96.041631	0.959712	0.942661	0.067159	Image_1.tif
2	12.568093	0.256809	0.289883	0	0.114786	0	1.830739	514	244.719844	698.120623	115.213203	0.974378	0.955390	0.045165	Image_1.tif

This table contains:

• Quantitative marker data (e.g., CD31, CK5, ANXA1, etc.)
• Morphometric properties (e.g., area, eccentricity, solidity, etc.)
• Cell metadata, such as coordinates (centroid-0 e.g x-axis, centroid-1 e.g y-axis) and file names (Filename).

Authors

Please cite this git page when using this tool! Miles Bailey
@milesbailey121