Pippeline is an interactive pipeline developed to standardize gene expression data processing in the NOWAC project.
After having installed this package in a regular manner (e.g., with install.packages() in R),
additional non-CTAN packages are needed to finalize the installation.
They can be installed by means of the included script. Write this in an R terminal:
source('install.R')
In case this does not succeed on a Linux box, run
sudo apt-get install libcairo2-dev libglu1-mesa-dev
in a Shell terminal. The former is to be able to install the Cairo package (from CTAN), a prerequisite for the arrayQualityMetrics package, and the latter is required by the lumi package. You may also have to run
install.packages('Cairo', dependencies=T)
in an R terminal afterwards.
The following assumes that this package is properly installed, and that you pointed R to the directory of this file.
First, load all necessary datasets. Example:
library(nowac)
You should now have a set of datasets available and can check that by:
data()
You may also have to configure the app. For this, edit the files shiny/options.csv and shiny/questionnaires.txt accordingly. The former contains the names of the R objects holding the available gene expression datasets (and their descriptions), while the latter contains the names of the R objects holding the available questionnaire data. See also the next section. Set working directory inside pippeline, if you are working from RStudio.
Also don't forget to set up pippelinen folder in globals.R - pipFolder. It is used for storing
results from pippeline, including SOP, outliers, reports and cleaned data.
That's it. Now write in an R terminal:
source('run.R')
You may be needed to generate outliers for the datasets. For this you can use SOP/nowaclean-gen.R script. This is interactive exploratory script you can use before you perform outlier removal step in the pippeline. It is not the actual part of pippeline.
A file with comma-separated values to store the available options and R objects in a spreadsheet-like manner. The app reads the file by means of read.csv(). The first row contains column labels and must not be changed. While the column values describe the dataset objects with (character) attributes like Design and Material, each entry for the up to three objects consists of two names, separated by a comma. The first name is interpreted as the name of the object holding the original dataset (a lumi object), the second is the (lumi) object with the negative controls ([gene expressions, negative controls]).
A file with one object name per line. The app reads them in by means of readLine().
- The browser window doesn't close after the Quit button is pressed.
There is currently no solution for this as JavaScript's
window.close()method is disabled in browsers due to security concerns. Just close the window with Ctrl/Cmd-W. - No PDFs are generated with R 3.2.3, rmarkdown 1.4, and pandoc 1.16.0.2, therefore the docFormat has been set to "html" in file globals.R. This may change with other versions; then the value "pdf" might be possible. This works in RStudio 1.0.136, though.
The following have contributed to the Pippeline development: Bjørn Fjukstad, Nikita Shvetsov, Therese H. Nøst, Hege Bøvelstad, Till Halbach, Einar Holsbø, Knut Hansen, Eiliv Lund, Lars Ailo Bongo