run,RnBClusterRun-method {RnBeads} | R Documentation |
Runs the analysis by submitting jobs for each module to the compute cluster
## S4 method for signature 'RnBClusterRun' run( object, analysis.id, config.xml, split.differential = TRUE, dry.run = FALSE, long.cmd.thres = 1024L, queue = NULL )
object |
|
analysis.id |
analysis id. used for naming submitted jobs and log files |
config.xml |
XML file specifying the analysis options and parameter settings |
split.differential |
flag indicating whether to split the differnetial methylation module into seperate jobs according to sample annotation column and region type. |
dry.run |
Prevent the actual job submission. Rather only write to a shell script file |
long.cmd.thres |
commands that are longer than this number will be encapsulated in shell scripts rather than being submitted as direct command |
queue |
The name of the queue the jobs are going to be submitted to |
Nothing of importance
Fabian Mueller
#specify the xml file for your analysis xml.file <- "MY_ANALYSIS_SETTINGS.XML" #set the cluster architecture specific to your environment arch <- new("ClusterArchitectureSGE") rnb.cr <- new("RnBClusterRun",arch) #set up the cluster so that 32GB of memory are required (SGE resource is called "mem_free") rnb.cr <- setModuleResourceRequirements(rnb.cr,c(mem_free="32G"),"all") #set up the cluster to use 4 cores on each node for all modules rnb.cr <- setModuleNumCores(rnb.cr,4L,"all") #set up the cluster to use 2 cores for the exploratory analysis module rnb.cr <- setModuleNumCores(rnb.cr,2L,"exploratory") #run the actual analysis (remove dry.run=TRUE, to really submit the jobs) run(rnb.cr, "rnbeads_analysis", xml.file, dry.run=TRUE)