Inspired by the seer from ancient Greek/Roman mythology, Laocoön, this package synthesizes different packages and methods available in the field of computer vision to create an automatic and efficient means of counting the number of cells in fluorescently stained Fucci cells.
To put it briefly, Fucci cells are stained with certain fluorescent proteins where a certain color corresponds with a certain part of the cell cycle. Comparing the proportion of cells in one part of the cell cycle to another part of the cell cycle can prove to be useful for clinical applications and can serve as an initial quantitative look at potentially diseased cells.
The only way that we can get these quantitative analyses is by actually counting the cells.
There can be a lot (and I mean a lot) of cells in a given image. Though there are existing
methods of automatic cell counting, these functionalities are restricted to plugins only
available in ImageJ, a Java-based image processing program
developed by the National Institute of Health. However, these methods are not used mainly because
they require tedious amounts of user input and are neither accurate nor efficient enough for
regular use (no shade).laocoon serves to address this problem by providing researchers
with a newer means of counting cells without staring at a screen for hours on end.
- Histogram equalization and counting the connected components.
- Gaussian filters and counting the regional maxima.
- Optional additions for quality control.
To download laocoon, you can either fork/download this repository or use PyPi via pip.
pip install laocoon
You must have mahotas, numpy, and pandas installed.
Input the absolute path to the folder containing your images, as well as the file type of
your images (acceptable image types are PNG, JPEG, JPG, and TIF). Each image must have
four different channels: DAPI, EdU, RFP, and GFP in the folder, and all similar images must
have the same prefix. Here's an example of how to execute the code after downloading it from pip install:
count_cells /Users/name/Desktop/fucci_images tif /Users/name/Downloads
The first argument is the absolute path to the folder containing the images, the second argumnet is the file type of the images, and the third is the absolute path to where the outputs should be saved.
If you are downloading it from GitHub, please navigate to the folder containing count_cells.py via your command
line, and then execute:
python count_cells.py /Users/name/Desktop/fucci_images tif /Users/name/Downloads
In executing this line of code, four methods of cell counting are performed on the images:
- Histogram equalization and Gaussian filter preprocessing, epsilon value quality control (saved as
hist_eps.csv) - Histogram equalization and Gaussian filter preprocessing (saved as
hist.csv) - Gaussian filter preprocessing and epsilon value quality control (saved as
eps.csv) - Just Gaussian filter preprocessing (
neither.csv)
The source code is divided into classes that represent the pipeline for counting the cells in each channel (DAPI, EdU, RFP, and GFP) and can be used individually, as well.
All files will be saved with a .csv extension in the directory that you have specified. The files will also contain calculations of proportions of cells in certain parts of the cell cycle, as well as the original counts from the different image channels.
This package is far from perfect! If you have any questions, comments, or suggestions, or if you're interested in contributing, please contact me at [email protected]. Response times may vary!