Added argument to WGP and HSP to allow more memory.

Upped the WGP VQSR memory to 32g to power through the filtering whole genome. TODO: Figure out what the right amount is.


git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5940 348d0f76-0448-11de-a6fe-93d51630548a

scala/qscript/playground/HybridSelectionPipeline.scala

@@ -36,21 +36,24 @@ class HybridSelectionPipeline extends QScript {
   @Argument(doc="the YAML file specifying inputs, interval lists, reference sequence, etc.", shortName="Y")
   var yamlFile: File = _
 
-  @Input(doc="level of parallelism for UnifiedGenotyper. By default is set to 20.", shortName="varScatter", required=false)
+  @Input(doc="level of parallelism for UnifiedGenotyper. By default set to 20.", shortName="varScatter", required=false)
   var variantCallerScatterCount = 20
 
-  @Argument(doc="memory limit for UnifiedGenotyper. By default is set to 4g.", shortName="varMemory", required=false)
+  @Argument(doc="memory limit for UnifiedGenotyper. By default set to 2g.", shortName="varMemory", required=false)
-  var variantCallerMemory = 4
+  var variantCallerMemory = 2
 
-  @Argument(doc="expand each target in input intervals by the specified number of bases (50 bases by default)", shortName="expand", required=false)
+  @Argument(doc="expand each target in input intervals by the specified number of bases. By default set to 50 bases.", shortName="expand", required=false)
   var expandIntervals = 50
 
+  @Argument(doc="pipeline memory limit. By default set to 2g.", shortName="pipeMemory", required=false)
+  var pipelineMemoryLimit = 2
+
   private var pipeline: Pipeline = _
 
   trait CommandLineGATKArgs extends CommandLineGATK {
     this.reference_sequence = qscript.pipeline.getProject.getReferenceFile
     this.intervals = List(qscript.pipeline.getProject.getIntervalList)
-    this.memoryLimit = 4
+    this.memoryLimit = pipelineMemoryLimit
   }
 
   def script() {

scala/qscript/playground/WholeGenomePipeline.scala

@@ -42,6 +42,9 @@ class WholeGenomePipeline extends QScript {
   @Argument(doc="Chunk size. Defaults to 3,000,000", shortName="chunk", required=false)
   var chunkSize = 3000000
 
+  @Argument(doc="Memory limit. Defaults to 4g", shortName="pipeMemory", required=false)
+  var pipelineMemoryLimit = 4
+
   val resources = "/humgen/gsa-pipeline/resources/5777/b37/"
   val reference = resources + "human_g1k_v37.fasta"
   val dbsnp = resources + "dbsnp_132.b37.vcf"
@@ -52,7 +55,7 @@ class WholeGenomePipeline extends QScript {
   trait CommandLineGATKArgs extends CommandLineGATK {
     this.reference_sequence = reference
     this.intervalsString = runIntervals
-    this.memoryLimit = 2
+    this.memoryLimit = pipelineMemoryLimit
   }
 
   case class Interval(chr: String, start: Long, stop: Long) {
@@ -214,6 +217,7 @@ class WholeGenomePipeline extends QScript {
     vr.tranches_file = projectBase + ".tranches"
     vr.recal_file = projectBase + ".recal"
     vr.jobOutputFile = vr.recal_file + ".out"
+    vr.memoryLimit = 32
     add(vr)
 
     for (tranche <- vr.TStranche) {
@@ -224,6 +228,7 @@ class WholeGenomePipeline extends QScript {
       ar.ts_filter_level = tranche.toDouble
       ar.out = projectBase + ".recalibrated." + tranche + ".vcf"
       ar.jobOutputFile = ar.out + ".out"
+      ar.memoryLimit = 32
       add(ar)
 
       val eval = new VariantEval with CommandLineGATKArgs
@@ -233,9 +238,10 @@ class WholeGenomePipeline extends QScript {
       eval.doNotUseAllStandardStratifications = true
       eval.doNotUseAllStandardModules = true
       eval.evalModule = List("SimpleMetricsByAC", "TiTvVariantEvaluator", "CountVariants")
-      eval.stratificationModule = List("EvalRod", "CompRod", "Novelty", "Filter", "FunctionalClass", "Sample")
+      eval.stratificationModule = List("EvalRod", "CompRod", "Novelty")
       eval.out = swapExt(ar.out, ".vcf", ".eval")
       eval.jobOutputFile = eval.out + ".out"
+      eval.memoryLimit = 32
       add(eval)
     }
   }