Analysis Types

An analysis is a discrete unit of computational work applied to a data set by a user within a running app. Scripts in the analysisTypes folder define a specific way of analyzing data that can be:

• executed as a synchronous or asynchronous job
• potentially reused by many apps

Here’s how to set an analysis type’s parameters and handler functions for use by runAnalyses and related modules.

All analysisTypes must provide, in folder analysisTypes/<typeGroup>/<analysisType>

• a config.yml file, with members (all are optional):
  • name = display name for the analysisType; default = <analysisType>
  • jobType = how/where the job should be executed; see job_execution.R; default = promise
  • options = job execution options, analogous to module settings; see existing prototypes. There are four UI columns for options display; use ‘type: empty’ for a blank position
  • packages = R or Bioconductor packages used by the analysisType; installed but not attached to running server process
  • classes = optional data classes used by the analysisType
  • modules = optional ui modules used by the analysisType
• a <analysisType>_methods.R file, with the following S3 generic functions:
  • setJobParameters.<analysisType> = convert reactives to static values for job promises
  • executeJob.<analysisType> = do the actual work of the job, potentially within a promise
  • load.<analysisType> = load the output of executeJob() for use by viewer modules
• a results viewer Shiny module used to populate the ‘results’ uiOutput in module viewResults via files:
  • <analysisType>_ui.R
  • <analysisType>_server.R

---