wiki:xQTLBiologistRun

Version 4 (modified by jvelde, 13 years ago) (diff)

--

TracNav(xQTL)?

xQTL workbench - Run QTL mapping

Start by clicking Run QTL mapping in the main menu. This will take you to the central screen for starting new or monitoring existing 'jobs', in this case QTL mapping.

Starting a new job

  • Click Create new job after selecting the compute location. Local means starting it on the same computer where the application is currently running. Cluster means submitting it to a remote location where more computational power might be available.

Step 1

  • Give a name to the output data matrix. If it happens to exist, a timestamp will be added to make it unique.
  • Select the kind of analysis you would like to run, e.g. Rqtl_analysis.
  • Select the amount of subjobs this computational task should be divided into. For larger sets of traits, it is beneficial to set this to higher value, so many jobs will run in parallel. However, if little power (CPUs) is available, a high setting may actually slow the task down. If you are in doubt, use the default of 5.

Step 2:

  • Select the datasets to use for this analysis. When many datasets are uploaded and tagged as pheno- and genotype data, be careful to select the right combination.
  • Select the parameters you would like to use.
  • Start the analysis by pressing Start, this will take you to the job manager where you can follow its progress.

Note that if you cannot find back your dataset, or are unhappy with any of the parameters, a bioinformatician (or admin) can configure them for you.

Monitoring job progression

  • In the Run QTL mapping screen, click on Go to job manager. This will take you to the job manager.
  • The status of all running subjobs can be seen here. The color coding indicates the status. Failed is red/-1, submitted orange/0, queued yellow/1, running blue/2, completed green/3.
  • Provenance is kept by storing all analysis parameters in the database. This can be viewed by holding your mouse cursor over the text marked as hover.
  • Provenance of completed jobs can be removed by clicking the delete button marked as . Note that this will not delete results already stored in the database.
  • Failed sub jobs (statuscode -1) can be resubmitted by clicking on them.
  • Note that you must have supplied your remote login credentials for resubmission of cluster jobs. (Needs a fix: currently this can only be done by submitting a new job)
  • The refresh page box can be used to speed up or slow down the polling interval. (5, 15, 30, 60 seconds or off)