gatk-3.8/scala/qscript/oneoffs/carneiro/dataProcessing.scala

229 lines
11 KiB
Scala
Executable File

import org.broadinstitute.sting.queue.extensions.gatk._
import org.broadinstitute.sting.queue.extensions.picard.PicardBamJarFunction
import org.broadinstitute.sting.queue.QScript
import org.broadinstitute.sting.queue.extensions.samtools.SamtoolsIndexFunction
import org.broadinstitute.sting.queue.function.ListWriterFunction
import scala.io.Source
class dataProcessing extends QScript {
qscript =>
@Input(doc="path to GATK jar", shortName="gatk", required=true)
var GATKjar: File = _
@Input(doc="path to AnalyzeCovariates.jar", shortName="ac", required=true)
var ACJar: File = _
// todo -- we should support the standard GATK arguments -R, -D [dbsnp],
// todo -- and indel files. Those should be defaulted to hg19 but providable on command line
@Input(doc="path to R resources folder inside Sting", shortName="r", required=true)
var R: String = _
@Input(doc="path to Picard FixMateInformation.jar. See http://picard.sourceforge.net/ .", shortName="fixmates", required=false)
var fixMatesJar: File = new java.io.File("/seq/software/picard/current/bin/FixMateInformation.jar")
@Input(doc="path to MarkDuplicates jar", shortName="dedup", required=false)
var dedupJar: File = new java.io.File("/seq/software/picard/current/bin/MarkDuplicates.jar")
@Input(doc="input BAM file", shortName="i", required=true)
var input: String = _
@Input(doc="final output BAM file base name", shortName="p", required=false)
var projectName: String = "combined"
@Input(doc="output path", shortName="outputDir", required=false)
var outputDir: String = ""
@Input(doc="the -L interval string to be used by GATK", shortName="L", required=false)
var intervalString: String = ""
// todo -- this shouldn't be allowed. We want a flag that says "output bams at intervals only" or not
@Input(doc="provide a .intervals file with the list of target intervals", shortName="intervals", required=false)
var intervals: File = new File("/humgen/1kg/processing/pipeline_test_bams/whole_genome_chunked.hg19.intervals")
// Reference sequence, dbsnps and RODs used by the pipeline
val reference: File = new File("/humgen/1kg/reference/human_g1k_v37.fasta")
val dbSNP: File = new File("/humgen/gsa-hpprojects/GATK/data/dbsnp_132_b37.leftAligned.vcf")
// TODO -- let's create a pre-merged single VCF and put it into /humgen/gsa-hpprojects/GATK/data please
val dindelPilotCalls: String = "/humgen/gsa-hpprojects/GATK/data/Comparisons/Unvalidated/1kg.pilot_release.merged.indels.sites.hg19.vcf"
val dindelAFRCalls: String = "/humgen/1kg/DCC/ftp/technical/working/20110126_dindel_august/AFR.dindel_august_release_merged_pilot1.20110126.sites.vcf.gz"
val dindelASNCalls: String = "/humgen/1kg/DCC/ftp/technical/working/20110126_dindel_august/ASN.dindel_august_release_merged_pilot1.20110126.sites.vcf.gz"
val dindelEURCalls: String = "/humgen/1kg/DCC/ftp/technical/working/20110126_dindel_august/EUR.dindel_august_release_merged_pilot1.20110126.sites.vcf.gz"
// Simple boolean definitions for code clarity
val knownsOnly: Boolean = true
val intermediate: Boolean = true
// General arguments to all programs
trait CommandLineGATKArgs extends CommandLineGATK {
this.jarFile = qscript.GATKjar
this.reference_sequence = reference
this.memoryLimit = Some(4)
this.isIntermediate = true
}
def script = {
var perLaneBamList: List[String] = Nil
var recalibratedBamList: List[File] = Nil
var recalibratedBamIndexList: List[File] = Nil
// Populates the list of per lane bam files to process (single bam or list of bams).
if (input.endsWith("bam")) {
perLaneBamList :+= input
}
else {
for (bam <- Source.fromFile(input).getLines())
perLaneBamList :+= bam
}
perLaneBamList.foreach { perLaneBam =>
// Helpful variables
val baseName: String = swapExt(new File(perLaneBam.substring(perLaneBam.lastIndexOf("/")+1)), ".bam", "").toString()
val baseDir: String = perLaneBam.substring(0, perLaneBam.lastIndexOf("/")+1)
// BAM files generated by the pipeline
val cleanedBam: String = baseName + ".cleaned.bam"
val fixedBam: String = baseName + ".cleaned.fixed.bam"
val dedupedBam: String = baseName + ".cleaned.fixed.dedup.bam"
val recalBam: String = baseName + ".cleaned.fixed.dedup.recal.bam"
// Accessory files
val targetIntervals: String = baseName + ".indel.intervals"
val metricsFile: String = baseName + ".metrics"
val preRecalFile: String = baseName + ".pre_recal.csv"
val postRecalFile: String = baseName + ".post_recal.csv"
val preOutPath: String = baseName + ".pre"
val postOutPath: String = baseName + ".post"
add(new target(perLaneBam, targetIntervals),
new clean(perLaneBam, targetIntervals, cleanedBam, knownsOnly), // todo -- use constrained movement mode to skip this
new fixMates(cleanedBam, fixedBam, intermediate), // todo -- use constrained movement mode to skip this
new dedup(fixedBam, dedupedBam, metricsFile), // todo -- generate index on fly here
new index(dedupedBam), // todo -- remove for on the fly index
new cov(dedupedBam, preRecalFile),
new recal(dedupedBam, preRecalFile, recalBam), // todo -- use GATK on the fly indexing?
new index(recalBam), // todo remove for on the fly indexing
new cov(recalBam, postRecalFile),
new analyzeCovariates(preRecalFile, preOutPath),
new analyzeCovariates(postRecalFile, postOutPath))
recalibratedBamList :+= new File(recalBam)
recalibratedBamIndexList :+= new File(recalBam + ".bai") // to hold next process by this dependency
}
// Helpful variables
val outName: String = qscript.projectName
val outDir: String = qscript.outputDir
// BAM files generated by the pipeline
val bamList: String = outDir + outName + ".list"
val cleanedBam: String = outDir + outName + ".cleaned.bam"
val fixedBam: String = outDir + outName + ".final.bam"
// Accessory files
val targetIntervals: String = outDir + outName + ".indel.intervals"
add(new writeList(recalibratedBamList, bamList, recalibratedBamIndexList),
new target(bamList, targetIntervals),
new clean(bamList, targetIntervals, cleanedBam, !knownsOnly), // todo -- use constrained movement mode to skip fix mates
new fixMates(cleanedBam, fixedBam, !intermediate)) // todo -- use constrained movement mode to skip this
}
class target (inBams: String, outIntervals: String) extends RealignerTargetCreator with CommandLineGATKArgs {
this.input_file :+= new File(inBams)
this.out = new File(outIntervals)
this.mismatchFraction = Some(0.0)
this.rodBind :+= RodBind("dbsnp", "VCF", dbSNP)
this.rodBind :+= RodBind("indels1", "VCF", dindelPilotCalls)
this.rodBind :+= RodBind("indels2", "VCF", dindelAFRCalls)
this.rodBind :+= RodBind("indels3", "VCF", dindelEURCalls)
this.rodBind :+= RodBind("indels4", "VCF", dindelASNCalls)
this.jobName = inBams + ".tgt"
if (!qscript.intervalString.isEmpty()) this.intervalsString :+= qscript.intervalString
else this.intervals :+= qscript.intervals
}
class clean (inBams: String, tIntervals: String, outBam: String, knownsOnly: Boolean) extends IndelRealigner with CommandLineGATKArgs {
this.input_file :+= new File(inBams)
this.targetIntervals = new File(tIntervals)
this.out = new File(outBam)
this.doNotUseSW = true
this.baq = Some(org.broadinstitute.sting.utils.baq.BAQ.CalculationMode.CALCULATE_AS_NECESSARY)
this.rodBind :+= RodBind("dbsnp", "VCF", dbSNP)
this.rodBind :+= RodBind("indels1", "VCF", dindelPilotCalls)
this.rodBind :+= RodBind("indels2", "VCF", dindelAFRCalls)
this.rodBind :+= RodBind("indels3", "VCF", dindelEURCalls)
this.rodBind :+= RodBind("indels4", "VCF", dindelASNCalls)
this.useOnlyKnownIndels = knownsOnly
this.sortInCoordinateOrderEvenThoughItIsHighlyUnsafe = true
this.jobName = inBams + ".clean"
if (!qscript.intervalString.isEmpty()) this.intervalsString ++= List(qscript.intervalString)
else this.intervals :+= qscript.intervals
}
class fixMates (inBam: String, outBam: String, intermediate: Boolean) extends PicardBamJarFunction {
@Input(doc="cleaned bam") var cleaned: File = new File(inBam)
@Output(doc="fixed bam") var fixed: File = new File(outBam)
override def inputBams = List(cleaned)
override def outputBam = fixed
this.jarFile = qscript.fixMatesJar
this.isIntermediate = intermediate
this.memoryLimit = Some(6)
this.jobName = inBam + ".fix"
}
class dedup (inBam: String, outBam: String, metricsFile: String) extends PicardBamJarFunction {
@Input(doc="fixed bam") var clean: File = new File(inBam)
@Output(doc="deduped bam") var deduped: File = new File(outBam)
override def inputBams = List(clean)
override def outputBam = deduped
override def commandLine = super.commandLine + " M=" + metricsFile
sortOrder = null
this.memoryLimit = Some(6)
this.jarFile = qscript.dedupJar
this.jobName = inBam + ".dedup"
}
// todo -- may we should use the picard version instead? What about telling all of the picard tools to
// todo -- generate BAM indices on the fly? That would be even better
class index (inBam: String) extends SamtoolsIndexFunction {
@Output(doc="bam index file") var outIndex: File = new File(inBam + ".bai")
this.bamFile = new File(inBam)
this.analysisName = inBam + ".index"
}
class cov (inBam: String, outRecalFile: String) extends CountCovariates with CommandLineGATKArgs {
this.rodBind :+= RodBind("dbsnp", "VCF", dbSNP)
this.covariate ++= List("ReadGroupCovariate", "QualityScoreCovariate", "CycleCovariate", "DinucCovariate")
this.input_file :+= new File(inBam)
this.recal_file = new File(outRecalFile)
}
class recal (inBam: String, inRecalFile: String, outBam: String) extends TableRecalibration with CommandLineGATKArgs {
this.input_file :+= new File (inBam)
this.recal_file = new File(inRecalFile)
this.out = new File(outBam)
}
class analyzeCovariates (inRecalFile: String, outPath: String) extends AnalyzeCovariates {
this.jarFile = qscript.ACJar
this.resources = qscript.R
this.recal_file = new File(inRecalFile)
this.output_dir = outPath
}
class writeList(inBams: List[File], outBamList: String, depIndices: List[File]) extends ListWriterFunction {
@Input(doc="bam indexes") var indexes: List[File] = depIndices // I need this dependency to hold creation of the list until all indices are done
this.inputFiles = inBams
this.listFile = new File(outBamList)
this.jobName = "bamList"
}
}