containerized_metabolomics

A virtual machine equipped with (mainly R-based) tools and packages for metabolomics analysis.

Background and rationale

Main focus is on an initial descriptive statistics module yielding p-values of differentail expression as well as Fold Change values; followed by pathway analyses.

Heavily builds on metaboanalyst and two example notebooks simply recreate demos at https://www.metaboanalyst.ca website.

The additional advantage of this container (besides easy deployment) is that it hosts R scripts wrapping some metaboanalyst pipelines together to provide a shortcut both working from a notebook or via command line.

Since most recent update, it is based on R 4.1 (after R 3.6).

Interactive notebook

To run the example notebook, simply pull the Docker container with

docker pull  szabogtamas/containerized_metabolomics:latest

Then start the RStudio server with something like:

docker run --rm -p 127.0.0.1:8787:8787 -v $PWD:/home/rstudio/local_files -e USERID=$UID -e PASSWORD=<yourpassword> szabogtamas/containerized_metabolomics

Your work will only get saved if it is in the local_files directory of the container (attached to the folder from where you ran the above command).
The recomended usage is to open a template notebook in repo_files (e.g.: repo_files/notebooks/script_capability_demo.Rmd) and save it in the local_files folder before starting to work on actual data.

Nextflow pipeline

To compile a standardized report from a batch of samples, a Nextflow pipeline is also available on the image. If the measured data is in the input_data folder of the current directory (alternatively, using test data in data_examples/pipeline_input_data), a basic report can be generated issuing the following command:

docker run -it -v $PWD:/home/rstudio/local_files \
  szabogtamas/containerized_metabolomics \
  nextflow run /home/rstudio/repo_files/nextflow/main.nf \
  --input_folder /home/rstudio/local_files/input_data \
  --report_filename "my_test_metabolomics_report.pdf" \
  --report_title "A test report" \
  --report_author "Me"

Additional tables will be in tables folder and individual figures in figures folder.

Please note that or some strange reason (and less frequently since last updates), Nextflow pipeline hangs sometimes when called from within the container. A more robust use case is when Nextflow is launched on a host or SGE and the individual processes are executed inside this container.

In the most recent rocker image, tinytex package manager is not working well and this results in an error when knitting the pdf report. Due to this obstacle, pdf figures are merged into a report using pdfunite, temporarily.

szabogtamas / containerized_metabolomics Goto Github PK