Added VQSR to the downsampling script and changed memory limits for the clean script.

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5817 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
carneiro 2011-05-18 20:07:42 +00:00
parent 4b00fd2688
commit b5b8cb959a
2 changed files with 38 additions and 4 deletions

View File

@ -3,6 +3,7 @@ package oneoffs.carneiro
import org.broadinstitute.sting.queue.QScript
import org.broadinstitute.sting.queue.extensions.gatk._
import scala.io.Source._
/**
* Created by IntelliJ IDEA.
* User: carneiro
@ -38,6 +39,12 @@ class downsampling extends QScript {
@Input(doc="Reference fasta file", shortName="R", required=false)
var reference: File = new File("/seq/references/Homo_sapiens_assembly19/v1/Homo_sapiens_assembly19.fasta")
@Input(doc="HapMap file", shortName="H", required=false)
var hapmap: File = new File("/humgen/gsa-hpprojects/GATK/data/Comparisons/Validated/HapMap/3.3/sites_r27_nr.b37_fwd.vcf")
@Input(doc="Omni file", shortName="O", required=false)
var omni: File = new File("/humgen/gsa-hpprojects/GATK/data/Comparisons/Validated/Omni2.5_chip/Omni25_sites_1525_samples.b37.vcf")
@Input(doc="dbSNP file", shortName="D", required=false)
var dbSNP: File = new File("/humgen/gsa-hpprojects/GATK/data/Comparisons/Validated/dbSNP/dbsnp_132_b37.leftAligned.vcf")
@ -55,6 +62,7 @@ class downsampling extends QScript {
val queueLogDir: String = ".qlog/"
val outFile: String = "cov.out"
val fullCoverageVCF = new File("/humgen/gsa-hpprojects/dev/carneiro/downsampling/analysis/fullcov/fullcov.F1.filtered.vcf")
val trancheTarget = "99.0"
def script = {
val nIntervals = math.min(200, countLines(targetIntervals))
@ -120,6 +128,32 @@ class downsampling extends QScript {
this.jobName = queueLogDir + outFile
}
// 3.) Variant Quality Score Recalibration - Generate Recalibration table
case class VQSR(inFile: File, tranchesFiles: File, outFile: File) extends VariantRecalibrator with CommandLineGATKArgs {
this.rodBind :+= RodBind("input", "VCF", inFile)
this.rodBind :+= RodBind("hapmap", "VCF", hapmap, "known=false,training=true,truth=true,prior=15.0")
this.rodBind :+= RodBind("omni", "VCF", omni, "known=false,training=true,truth=true,prior=12.0")
this.rodBind :+= RodBind("dbsnp", "VCF", dbSNP, "known=true,training=false,truth=false,prior=10.0")
this.use_annotation ++= List("QD", "HaplotypeScore", "MQRankSum", "ReadPosRankSum", "HRun")
this.tranches_file = tranchesFile
this.recal_file = outFile
this.allPoly = true
this.tranche ++= List("100.0", "99.9", "99.5", "99.3", "99.0", "98.9", "98.8", "98.5", "98.4", "98.3", "98.2", "98.1", "98.0", "97.9", "97.8", "97.5", "97.0", "95.0", "90.0")
this.analysisName = t.name + "_VQSR"
this.jobName = queueLogDir + outFile
}
// 4.) Apply the recalibration table to the appropriate tranches
case class applyVQSR (inFile: File, tranchesFiles: File, outFile: File) extends ApplyRecalibration with CommandLineGATKArgs {
this.rodBind :+= RodBind("input", "VCF", inFile)
this.tranches_file = tranchesFile
this.recal_file = inFile
this.ts_filter_level = trancheTarget
this.out = outFile
this.analysisName = outFile + "_AVQSR"
this.jobName = queueLogDir + outFile
}
case class eval (inFile: File, outFile: File) extends VariantEval with CommandLineGATKArgs {
this.noST = true
this.noEV = true

View File

@ -1,6 +1,3 @@
import org.broadinstitute.sting.queue.QScript
import org.broadinstitute.sting.queue.extensions.gatk.{RealignerTargetCreator, RodBind, IndelRealigner}
/**
* Created by IntelliJ IDEA.
* User: carneiro
@ -9,6 +6,9 @@ import org.broadinstitute.sting.queue.extensions.gatk.{RealignerTargetCreator, R
* To change this template use File | Settings | File Templates.
*/
import org.broadinstitute.sting.queue.extensions.gatk.{IndelRealigner, RealignerTargetCreator, RodBind}
import org.broadinstitute.sting.queue.QScript
class justClean extends QScript {
@ -62,7 +62,7 @@ class justClean extends QScript {
clean.doNotUseSW = false
clean.jobName = queueLogDir + outBam + ".clean"
clean.jarFile = GATKjar
clean.memoryLimit = 24
clean.memoryLimit = 8
clean.scatterCount = 84
add(target, clean);