Optimize XCMS analytical parameters by running IPO workflow (Peak picking + Retention time correction and grouping optimization) on Metabolights studies within the PhenoMeNal EU project.
Runs the IPO tool that includes as a 2 steps workflow the peak picking step first and then retention time correction and peak grouping optimization steps together, using wft4galaxy api. The workflow can process one (scripts/run_ipo_workflow.R) or several (scripts/run_ipo_on_several_studies.sh) Metabolights studies (http://www.ebi.ac.uk/metabolights/) using wft4galaxy (http://wft4galaxy.readthedocs.io) as an API accessing a Galaxy instance with an installed PhenoMeNal e-infrastructure. This script processes all assays separately. It picks 5% but at least 10 data files to save processing time. For the grouping step, the sample metadata file is necessary in order to take into account the experimental design factors. They are integrated into the best xcmsSet R object resulting from the peak picking step.
**Important**: The workflow consists in optimizing a subset of parameters at once only, to reduce computation time. You need to download the workflow optimizing the parameters of your choice from Galaxy (into a .ga file). The workflow present in template/ipo_workflow_peakwidth_bw.ga
-
Docker
-
Make sure that wft4galaxy-docker is available on your machine.
Visit http://wft4galaxy.readthedocs.io/installation.html#id2 for installation instructions. -
Install isa2w4m container from PhenoMeNal (Convert ISA-tab into 3 files: sample metadata, variable metadata, sample x variable matrix)
$ docker pull container-registry.phenomenal-h2020.eu/phnmnl/isa2w4m -
(optionnal) You can install the Metabolights Downloader container from PhenoMeNal (Pierrick Roger - CEA)
$ docker pull container-registry.phenomenal-h2020.eu/phnmnl/mtbls-dwnld
to download a Metabolights study:
$ docker run -v /home/user/workdir/studies/:/studies container-registry.phenomenal-h2020.eu/phnmnl/mtbls-dwnld -a -o /studies MTBLS213
This step is done automatically if you run the workflow on several studies. See the section at the end
-
-
The R packages Risa and R.utils need to be installed.
-
install.packages("R.utils") # Risa source("https://bioconductor.org/biocLite.R") biocLite("Risa")
-
Prepare the working environment.
git clone https://github.com/gabrielctn/phnmnl_IPO_workflow.git
cd phnmnl_IPO_workflow
mkdir output
chmod a+x ./scripts/run_IPO_workflow.r The working directory should look like:
work_dir
|_ output (to build)
|_ studies
| |_ MTBLS213
| |_ ...
|_ template
| |_ ipo_workflow_peakwidth_bw.ga
|_ python
| |_ simple_comp.py
|_ scripts
|_ run_ipo_workflow.R
|_ run_ipo_on_several_studies.sh
Run the script without any argument to display help message.
./run_ipo_workflow.r
Arguments
study_path - Local path containing the MTBLS study
study_name - Name of the study to be processed (e.g. MTBLS433)
ga_file_template - Galaxy workflow file (.ga) for the IPO workflow
output_path - Local path used to store the result files
log_file - Path to csv file used for log messages
galaxy_url - URL to Galaxy server on which to run the workflow
galaxy_key - API key to access the Galaxy server
debug - Get debug output from wft4galaxy (true, false) default: false
logger - Enable logger for wft4galaxy (true, false) default: false
./run_ipo_workflow.r
study_path="studies"
study_name="MTBLS213"
ga_file_template="template/ipo_workflow_peakwidth_bw.ga"
output_path="output"
log_file="MTBLS213_logs.csv"
galaxy_url="http://192.168.99.100:30700"
galaxy_key="ce30966564d6a42b22f951ca023081ed"Same requirements as described previously.
This script runs the script described above (run_ipo_workflow.R) over all the Metabolights studies mentioned in a given folder ("studies" for example). The difference lies in the fact that the studies are downloaded automatically, only the Metabolights ID in a simple text file are required.
Usage: run_ipo_on_several_studies.sh [options]
Repo: https://github.com/gabrielctn/phnmnl_IPO_workflow
Runs the IPO workflow held in PhenoMeNal over all Metabolights studies mentionned in the text file given in argument with wft4galaxy.
The workflow consists in 2 steps:
1. Peakpicking
2. Grouping and retention time correction
Options:
-h, --help Print this help message.
-r, --script PATH Path to 'run_ipo_workflow.R' script.
-s, --studies PATH Path to the text file containing all the Metabolights studies. 1 line = 1 MTBLSXXX.
-t, --template PATH Path to the galaxy workflow (.ga) template.
-o, --output PATH Path to the output folder, where results will be created.
-u, --galaxyurl URL The URL for the Galaxy server. For example: http://192.168.99.100:30700/
-k, --galaxykey KEY The galaxy api key of the actual galaxy server. Can be found in the settings of the admin galaxy interface. For example: 822c174c483b7a3675c313c1a4466f12
-d, --debug true/false Print debug infos.
-l, --logger true/false Print log infos.
Since the purpose of this script is to run the workflow on different studies, outputs will not follow a specific template. That is why the API does not compare the resulting files with the "expected" ones.
----------------------------------------------------------------------
Ran 1 test in 1600.722s
OK
2018-05-12 19:30:31,667 [wft4galaxy.app.runner] [DEBUG] wft4galaxy.run_tests exiting with code: 0
The results will appear in the registered histories on Galaxy (browser) and on the "results" folder locally, in the folder "output".
Example with MTBLS217:
# assay NEG
$ ls output/MTBLS117/a_mtbls117_DIMS_NEG_mass_spectrometry/results/ipo/
best_xcmsSet.RData ipo4xcmsSet.log.txt IPO_parameters4xcmsSet.tsv run_instrument_infos.tsv
ipo4retcor.log.txt IPO_parameters4retcorGroup.tsv resultPeakpicking.RData
# assay POS
$ ls output/MTBLS117/a_mtbls117_DIMS_POS_mass_spectrometry/results/ipo/
best_xcmsSet.RData ipo4xcmsSet.log.txt IPO_parameters4xcmsSet.tsv run_instrument_infos.tsv
ipo4retcor.log.txt IPO_parameters4retcorGroup.tsv resultPeakpicking.RData