Cell Viability Assay

This repository contains the source code for analyzing cell viability assays using high-content screening (HCS) images. The pipeline includes image processing and data mining for quantitative analysis of the number of cells in each image.

Images

The images were generated from a phenotypic assay of VERO CCL81 cells provided by Dr. Rafael E. Marques (LNBio), which were infected with the Mayaro virus and treated with the vehicle DMSO, alongside non-infected controls. The cells were stained with the fluorophore Hoechst 33342, a DNA marker, and the images were acquired using the Operetta microscope (10x magnification).

Analysis

The images were processed using open-source tools, such as CellProfiler, for image processing analysis, and then data mined in Python for quantitative analysis. The custom CellProfiler pipeline included pre-processing, segmentation of nuclei, and calculation of metrics (number of cells).

To run the pipeline, follow the instructions below:

1. Install CellProfiler and the required plugins.

2. Run the pipeline and data mining using the run.sh script:

# Running on HPC marvin machine
sbatch run.sh -m marvin

or

# Running on local machine
bash run.sh -m local -p /path/to/CellProfiler-plugins/active_plugins

The output files will be saved in the results directory and will include the following files:

results/
├── summary.csv                   # Summary of the number of cells per well
├── Experiment.csv                # Raw data of the number of cells per image (CellProfiler output)
├── Nuclei.csv                    # Raw data of the segmented nuclei per image (CellProfiler output)
├── Image.csv                     # Raw information per image (CellProfiler output)
└── plate_map/
    └── number_of_cells.html      # Interactive visualization of the number of cells per well

Benchmarking

The results obtained using the cell viability protocols developed with the open-source software CellProfiler were benchmarked against the reference protocol applied in the proprietary high-content image analysis software, Columbus.

To reproduce the benchmarking analysis, run the jupyter notebook:

jupyter notebook

Then, open the notebook Cell Viability Analysis - Comparing CellProfiler and Columbus.ipynb in the benchmarking directory. Follow the instructions in the notebook to compare the results obtained using CellProfiler and Columbus.

Authors

• Daniel C. Vieira
• João V. S. Guerra
• José G. C. Pereira
• Kayllany L. S. Oliveira

License

This software is licensed under the terms of the GNU General Public License version 3 (GPL3) and is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.