zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

now processes either a single bam file or a list of bam files in parallel. 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5256 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
carneiro 2011-02-17 02:07:22 +00:00
parent 52eedaf22d
commit b4da843c49
1 changed files with 96 additions and 80 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
scala/qscript/oneoffs/carneiro/dataProcessing.scala
View File

@ -2,6 +2,7 @@ import org.broadinstitute.sting.queue.extensions.gatk._
import org.broadinstitute.sting.queue.extensions.picard.PicardBamJarFunction import org.broadinstitute.sting.queue.extensions.picard.PicardBamJarFunction
import org.broadinstitute.sting.queue.QScript import org.broadinstitute.sting.queue.QScript
import org.broadinstitute.sting.queue.extensions.samtools.SamtoolsIndexFunction import org.broadinstitute.sting.queue.extensions.samtools.SamtoolsIndexFunction
import scala.io.Source
class dataProcessing extends QScript { class dataProcessing extends QScript {
@ -14,7 +15,7 @@ class dataProcessing extends QScript {
var dedupJar: File = _ var dedupJar: File = _
@Input(doc="input BAM file", shortName="input", required=true) @Input(doc="input BAM file", shortName="input", required=true)
var inputBam: File = _ var inputBam: String = _
@Input(doc="output path", shortName="outputDir", required=false) @Input(doc="output path", shortName="outputDir", required=false)
var outputDir: String = "" var outputDir: String = ""
@ -31,11 +32,25 @@ class dataProcessing extends QScript {
def script = { def script = {
// Populates the list of files to process either from a single bam or from a list of bams.
var bamList: List[String] = Nil
if (inputBam.endsWith("bam")) {
bamList :+= inputBam
}
else { // todo: maybe check if the filename ends with .list?
for (bam <- Source.fromFile(inputBam).getLines())
bamList :+= bam
}
bamList.foreach { bam =>
println("DEBUG: bam = " + new File(bam))
// Files generated by the pipeline // Files generated by the pipeline
val baseName: String = swapExt(qscript.inputBam, ".bam", "").toString() val baseName: String = swapExt(new File(bam.substring(bam.lastIndexOf("/")+1)), ".bam", "").toString()
val cleanedBam: String = baseName + ".cleaned.bam" val cleanedBam: String = baseName + ".cleaned.bam"
val dedupedBam: String = baseName + ".cleaned.dedup.bam" val dedupedBam: String = baseName + ".cleaned.dedup.bam"
val metricsFile: String = swapExt(qscript.inputBam, "bam", "metrics").toString() val metricsFile: String = baseName + ".metrics"
val targetIntervals: String = baseName + ".indel.intervals" val targetIntervals: String = baseName + ".indel.intervals"
val recalFile: String = baseName + ".recal.csv" val recalFile: String = baseName + ".recal.csv"
val recalBam: String = baseName + ".cleaned.dedup.recal.bam" val recalBam: String = baseName + ".cleaned.dedup.recal.bam"
@ -58,7 +73,7 @@ class dataProcessing extends QScript {
val target = new RealignerTargetCreator with CommandLineGATKArgs val target = new RealignerTargetCreator with CommandLineGATKArgs
target.input_file :+= qscript.inputBam target.input_file :+= new File(bam)
target.out = new File(targetIntervals) target.out = new File(targetIntervals)
target.mismatchFraction = Some(0.0) target.mismatchFraction = Some(0.0)
target.rodBind :+= RodBind("dbsnp", "VCF", dbSNP) target.rodBind :+= RodBind("dbsnp", "VCF", dbSNP)
@ -73,7 +88,7 @@ class dataProcessing extends QScript {
// 2.) Clean without SW // 2.) Clean without SW
val clean = new IndelRealigner with CommandLineGATKArgs val clean = new IndelRealigner with CommandLineGATKArgs
clean.input_file :+= qscript.inputBam clean.input_file :+= new File(bam)
clean.targetIntervals = new File(targetIntervals) clean.targetIntervals = new File(targetIntervals)
clean.out = new File(cleanedBam) clean.out = new File(cleanedBam)
clean.doNotUseSW = true clean.doNotUseSW = true
@ -118,4 +133,5 @@ class dataProcessing extends QScript {
add(target, clean, dedup, index, cov, recal) add(target, clean, dedup, index, cov, recal)
} }
}
} }