R functions allowing for the assessment of colocalization in a High-Througput fashion.
The provided piece of code was tested on both Linux (Ubuntu 16.04) and Windows (Windows 7) distributions. It has not been tested on Mac OS but should theoretically work. In any case, you will need to have the proper Java version (64 or 32 bits according to your machine) installed on your computer in the aim to have rJava package added to R.
For Linux distributions, you might need to install additionnal packages such as fftw3 so that you can install the R tiff package:
- sudo apt-get install libfftw3-dev libfftw3-doc
Before using the provided function you will obviously need to have R installed:
I advise you to use a GUI, such as RStudio, to make its use simpler:
You will need to have a few packages installed. For doing so, open R and run these lines in the console:
- biocLite(pkgs=c('EBImage','flowCore'), ask=F)
- install.packages(pkgs = c('shiny','tiff','reshape','RODBC','foreach','doParallel','stringi','naturalsort','rChoiceDialogs'))
or simply source the InstallPackages.R after opening it with R.
Open RGui/RStudio and open the provided scripts
ColocPixelAnalysis_Func.R (in the
Functions folder). After sourcing each of them, you can type in the console :
- coloc.HTS(getImInfo(SQL.use = F))
A first window will open to select the location of the images you want to analyze. After selecting the folder in which results will be exported in a second window, the analysis willl start. If you want to see a more detailed description of the functions and their arguments, please refer to
GetImInfo.R will only run with a specific nomenclature regarding images names.
- First, the images will need to be in *.tiff file format.
- The names should be given according to the well, site, and channel that have been acquired : ExpName_WellName_sSiteNumber_wChannelNumber
For example, if you have acquired an image from the well A01 of a plate, on site 1, using the first channel of your microscope, its name should be
- If you wish to analyze a timecourse experiment, images from each timepoint should be stored in a folder named TimePoint_Time, with the same nomenclature that described previously.
For example, if you ran an experiment with 10 timepoints, images shall be stored in folders named
TimePoint_10. Obviously, each batch of images in each folder will have the same name.
Inside the choosen results location, you should find 2 files and a serie of subfolders in a folder called "Results".
The first file, Results.csv, is a comma-delimited table in which calculated colocalization parameters are recorded for each acquired site. A more detailed description of these parameters can be consulted in
The second, namely Boxplot.pdf, is an informative boxplot depicting summarized results well by well (grouping sites altogether)
Every subfolder (one per timepoint) contains directories named according to the wells that were acquired. Each well directory contains as many pdfs as they ares sites inside the well, and one image. The pdf shows the bi-parametric plot of pixel intensities in assesed channels, together with the associated colocalization parameters. The recorder image is a color-combined image from site 1, in which detected Green and Red signals are detected. This image can be useful to refine segmentation parameters (see
Once installed, the suite can be tested on the provided test images in the
Testing-images folder. Please refer to
Images info.txt for more information about these images. After sourcing the R scripts, type in the console :
- coloc.HTS(getImInfo(SQL.use = F), Cyto = 'Red')
And select the folder where you stored the previously downloaded images.