Reproducing the FrenchFISH manuscript analyses

This repository contains the code required to reproduce the results and figures of the FrenchFISH manuscript. FrenchFISH is a method and corresponding R package for for correcting spot counts used for quantifying DNA copy-number from fluorescence in situ hybridisation of tissue sections. It can correct either spot counts from manual counting, or from an automatic spot counting software. The underlying models used to correct these counts are Poissonian.

Installing the frenchFISH R package

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("frenchFISH")

Running analyses and generating figures

After installing the frenchFISH package, the analyses and figures of the paper can be generated by running frenchFISH_analyses.Rmd. Before running, it is necessary that path_to_dir in the markdown file be set to the path to frenchFISH_analyses. In addition to simulating FISH spot count data for correction, this markdown file will access the automatic_counts and manual_counts directories, where CSV files containing the uncorrected spot counts for 12 ovarian cancer FFPE-tissue sections are stored.

Downloading the FISH images used in the paper

12 ovarian cancer FFPE-tissue sections assessed in the paper can be downloaded from the Cell Image Library. These images comprise z-stacks of TIFF images, where each z-stack is one of the tiles of the tissue section images used for analysis.

Regenerating spot count CSV files

Once the 12 image directories have been downloaded, they should be added to a directory called image_processing/input_data. All images from each tissue section should be included in a subdirectory named as that tissue section is named. image_processing/input_data should therefore include 12 subdirectories containing all TIFF images:

• image_processing/input_data/BL_024216_Myc_Terc
• image_processing/input_data/BL024199_Myc_Terc
• image_processing/input_data/BL32077_Myc_Terc
• image_processing/input_data/BL32080_Myc_Terc
• image_processing/input_data/PS09_20676_2B_Myc_Terc
• image_processing/input_data/PS09_287383C_Myc_Terc
• image_processing/input_data/PS11_10021_2B_Myc_Terc
• image_processing/input_data/PS11_167511L_Myc_Terc
• image_processing/input_data/SC_007
• image_processing/input_data/SC_011
• image_processing/input_data/SC_028
• image_processing/input_data/SC_030

image_processing contains the necessary scripts to reproduce the uncorrected spot count CSVs used above. Before proceeding, create an empty directory image_processing/results for storing image processing results. First, run image_processing/run_focus_selector.sh, which calls image_processing/select_focus_project_x.ijm on each of the 12 directories to generate JPEG max projections of the 10 layers around the most focused layer in each TIFF z-stack. Finally, run image_processing/image_processing.R to automatically count the spots in these max projection images using the FishalyzeR package.

Disclaimer

Note that this is prerelease software. Please use accordingly.