Headless mzMine Batch Runs
mzMine users are used to running large analyses via the Batch mode. Once the analysis steps are selected in the 'Batch queue editor' they can be saved to a file, which can be run on the command line. Since running the 'mzMine batch mode' requires the presence of the Graphical User Interface (GUI) resulting in poor scalability of having to use one of the GUI machines or an OnDemand session, which are not designed to run large-scale analyses, but are there for light code editing or viewing of the results it is preferable to run large-scale mzMine analyses via the command-line inside batch jobs scheduled on HiPerGator. However, doing the latter requires a user to set correct paths to the input files and adjust their cluster job script to run the batch script from the correct location possibly every time a new job is run even though the batch parameters may not changed at all the above approach may be taxing. To lower the barrier of entry to running mzMine on HiPerGator we have collaborated with the SECIM project to develop scripts that make running Running MZMine on HiPerGator in 'Headless mode easier and more automated. Incidentally, this allows external programs or pipelines to drive mzMine analyses without any human intervention if all pre-requisites are met. Following is the draft documentation for headless mzMine execution on HiPerGator. Please feel free to contact us if you have any questions or use UFRC support site to submit any questions, enhancement requests, or parameter file additions.
- the ‘$’ symbol denotes the command-line (shell) prompt
Create directories for mzmine analyses of positive and negative data
Make the input data files (e.g. *.mzML files) available for MZmine. Separate the positive and negative sample into different input directories.
Job Script and Job Submission
Submit analyses for positive and negative datasets separately passing the appropriate input directory path to mzmine_run_batch. The defaults are set for a medium-sized project.
Run the following command to see all available options
$ mzmine_run_batch -h
Required arguments: -s <see the list shown by mzmine_run_batch -h as the templates might have changed> -p pos, or neg
-i input directory with the positive or negative datasets
To run MZmine on (1) positive data using (2) a lipid batch file,run the following command:
$ mzmine_run_batch -i test1/mzmine_pos -p pos -s lipid
To run MZmine on negative data using a lipid batch file, run the following command:
$ mzmine_run_batch -i test1/mzmine_neg -p neg -s lipid
The mzmine_run_batch will submit the job unless you specify the '-n' argument
-n, --dryrun Dry run. Prepare xml and job files, but do not submit the job