Added protected artifact for gatk extensions

protected/biopet-gatk-extensions/.gitignore

+/target/
\ No newline at end of file

protected/biopet-gatk-extensions/pom.xml

+<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+    <modelVersion>4.0.0</modelVersion>
+
+    <groupId>nl.lumc.sasc</groupId>
+    <artifactId>BiopetGatkExtensions</artifactId>
+    <version>0.2.0-DEV</version>
+    <packaging>jar</packaging>
+    
+    <inceptionYear>2014</inceptionYear>
+    <name>BiopetGatkExtensions</name>
+    <url>http://maven.apache.org</url>
+
+    <properties>
+        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
+        <sting.unpack.phase>prepare-package</sting.unpack.phase>
+        <sting.shade.phase>package</sting.shade.phase>
+        <app.main.class>nl.lumc.sasc.biopet.core.BiopetExecutable</app.main.class>
+    </properties>
+
+    <dependencies>
+        <dependency>
+            <groupId>nl.lumc.sasc</groupId>
+            <artifactId>BiopetFramework</artifactId>
+            <version>0.2.0-DEV</version>
+        </dependency>
+        <dependency>
+            <groupId>org.broadinstitute.gatk</groupId>
+            <artifactId>gatk-queue-extensions-distribution</artifactId>
+            <version>3.3</version>
+        </dependency>
+    </dependencies>
+    <build>
+        <resources>
+            <resource>
+                <directory>src/main/resources</directory>
+                <includes>
+                    <include>**/*</include>
+                </includes>
+            </resource>
+        </resources>
+        <plugins>
+            <plugin>
+                <groupId>org.scala-tools</groupId>
+                <artifactId>maven-scala-plugin</artifactId>
+                <version>2.15.2</version>
+                <executions>
+                    <execution>
+                        <id>scala-compile</id>
+                        <goals>
+                            <goal>compile</goal>
+                            <goal>testCompile</goal>
+                        </goals>
+                        <configuration>
+                            <args>
+                                <arg>-dependencyfile</arg>
+                                <arg>${project.build.directory}/.scala_dependencies</arg>
+                                <arg>-deprecation</arg>
+                                <arg>-feature</arg>
+                            </args>
+                        </configuration>
+                    </execution>
+                </executions>
+            </plugin>
+            <plugin>
+                <groupId>org.apache.maven.plugins</groupId>
+                <artifactId>maven-jar-plugin</artifactId>
+                <version>2.5</version>
+                <configuration>
+                    <archive>
+                        <manifest>
+                            <addDefaultImplementationEntries>true</addDefaultImplementationEntries>
+                            <addDefaultSpecificationEntries>true</addDefaultSpecificationEntries>
+                        </manifest>
+                    </archive>
+                </configuration>
+            </plugin>
+            <plugin>
+                <groupId>org.apache.maven.plugins</groupId>
+                <artifactId>maven-compiler-plugin</artifactId>
+                <version>2.3.2</version>
+                <configuration>
+                    <showDeprecation>true</showDeprecation>
+                </configuration>
+            </plugin>
+        </plugins>
+    </build>
+</project>

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/AnalyzeCovariates.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class AnalyzeCovariates(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.AnalyzeCovariates with GatkGeneral {
+}
+
+object AnalyzeCovariates {
+  def apply(root: Configurable, before: File, after: File, plots: File): AnalyzeCovariates = {
+    val ac = new AnalyzeCovariates(root)
+    ac.before = before
+    ac.after = after
+    ac.plots = plots
+    return ac
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/ApplyRecalibration.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class ApplyRecalibration(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.ApplyRecalibration with GatkGeneral {
+  override def afterGraph {
+    super.afterGraph
+
+    if (config.contains("scattercount")) scatterCount = config("scattercount")
+
+    nt = Option(getThreads(3))
+    memoryLimit = Option(nt.getOrElse(1) * 2)
+    ts_filter_level = config("ts_filter_level")
+  }
+}
+
+object ApplyRecalibration {
+  def apply(root: Configurable, input: File, output: File, recal_file: File, tranches_file: File, indel: Boolean = false): ApplyRecalibration = {
+    val ar = if (indel) new ApplyRecalibration(root) {
+      mode = org.broadinstitute.gatk.tools.walkers.variantrecalibration.VariantRecalibratorArgumentCollection.Mode.INDEL
+      defaults ++= Map("ts_filter_level" -> 99.0)
+    }
+    else new ApplyRecalibration(root) {
+      mode = org.broadinstitute.gatk.tools.walkers.variantrecalibration.VariantRecalibratorArgumentCollection.Mode.SNP
+      defaults ++= Map("ts_filter_level" -> 99.5)
+    }
+    ar.input :+= input
+    ar.recal_file = recal_file
+    ar.tranches_file = tranches_file
+    ar.out = output
+    return ar
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/BaseRecalibrator.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class BaseRecalibrator(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.BaseRecalibrator with GatkGeneral {
+  memoryLimit = Option(4)
+  override val defaultVmem = "8G"
+
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+  if (config.contains("dbsnp")) knownSites :+= new File(config("dbsnp").asString)
+  if (config.contains("known_sites")) knownSites :+= new File(config("known_sites").asString)
+}
+
+object BaseRecalibrator {
+  def apply(root: Configurable, input: File, output: File): BaseRecalibrator = {
+    val br = new BaseRecalibrator(root)
+    br.input_file :+= input
+    br.out = output
+    br.afterGraph
+    return br
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/CombineGVCFs.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class CombineGVCFs(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.CombineGVCFs with GatkGeneral {
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+}
+
+object CombineGVCFs {
+  def apply(root: Configurable, input: List[File], output: File): CombineGVCFs = {
+    val cg = new CombineGVCFs(root)
+    cg.variant = input
+    cg.o = output
+    return cg
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/CombineVariants.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class CombineVariants(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.CombineVariants with GatkGeneral {
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+}
+
+object CombineVariants {
+  def apply(root: Configurable, input: List[File], output: File): CombineVariants = {
+    val cv = new CombineVariants(root)
+    cv.variant = input
+    cv.out = output
+    return cv
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/GatkGeneral.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import nl.lumc.sasc.biopet.core.BiopetJavaCommandLineFunction
+import org.broadinstitute.gatk.queue.extensions.gatk.CommandLineGATK
+
+trait GatkGeneral extends CommandLineGATK with BiopetJavaCommandLineFunction {
+  memoryLimit = Option(3)
+
+  if (config.contains("gatk_jar")) jarFile = config("gatk_jar")
+
+  override val defaultVmem = "7G"
+
+  if (config.contains("intervals", submodule = "gatk")) intervals = config("intervals", submodule = "gatk").asFileList
+  if (config.contains("exclude_intervals", submodule = "gatk")) excludeIntervals = config("exclude_intervals", submodule = "gatk").asFileList
+  reference_sequence = config("reference", submodule = "gatk")
+  gatk_key = config("gatk_key", submodule = "gatk")
+  if (config.contains("pedigree", submodule = "gatk")) pedigree = config("pedigree", submodule = "gatk").asFileList
+}

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/GenotypeGVCFs.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class GenotypeGVCFs(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.GenotypeGVCFs with GatkGeneral {
+  annotation ++= config("annotation", default = Seq("FisherStrand", "QualByDepth", "ChromosomeCounts")).asStringList
+
+  if (config.contains("dbsnp")) dbsnp = config("dbsnp")
+  if (config.contains("scattercount", "genotypegvcfs")) scatterCount = config("scattercount")
+
+  if (config("inputtype", default = "dna").asString == "rna") {
+    stand_call_conf = config("stand_call_conf", default = 20)
+    stand_emit_conf = config("stand_emit_conf", default = 0)
+  } else {
+    stand_call_conf = config("stand_call_conf", default = 30)
+    stand_emit_conf = config("stand_emit_conf", default = 0)
+  }
+}
+
+object GenotypeGVCFs {
+  def apply(root: Configurable, gvcfFiles: List[File], output: File): GenotypeGVCFs = {
+    val gg = new GenotypeGVCFs(root)
+    gg.variant = gvcfFiles
+    gg.out = output
+    return gg
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/HaplotypeCaller.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import nl.lumc.sasc.biopet.core.config.Configurable
+import org.broadinstitute.gatk.utils.variant.GATKVCFIndexType
+
+class HaplotypeCaller(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.HaplotypeCaller with GatkGeneral {
+  override def afterGraph {
+    super.afterGraph
+
+    min_mapping_quality_score = config("minMappingQualityScore", default = 20)
+    if (config.contains("scattercount")) scatterCount = config("scattercount")
+    if (config.contains("dbsnp")) this.dbsnp = config("dbsnp")
+    this.sample_ploidy = config("ploidy")
+    nct = config("threads", default = 1)
+    bamOutput = config("bamOutput")
+    memoryLimit = Option(nct.getOrElse(1) * 2)
+    if (config.contains("allSitePLs")) this.allSitePLs = config("allSitePLs")
+    if (config.contains("output_mode")) {
+      import org.broadinstitute.gatk.tools.walkers.genotyper.OutputMode._
+      config("output_mode").asString match {
+        case "EMIT_ALL_CONFIDENT_SITES" => output_mode = EMIT_ALL_CONFIDENT_SITES
+        case "EMIT_ALL_SITES"           => output_mode = EMIT_ALL_SITES
+        case "EMIT_VARIANTS_ONLY"       => output_mode = EMIT_VARIANTS_ONLY
+        case e                          => logger.warn("output mode '" + e + "' does not exist")
+      }
+    }
+
+    if (config("inputtype", default = "dna").asString == "rna") {
+      dontUseSoftClippedBases = config("dontusesoftclippedbases", default = true)
+      stand_call_conf = config("stand_call_conf", default = 5)
+      stand_emit_conf = config("stand_emit_conf", default = 0)
+    } else {
+      dontUseSoftClippedBases = config("dontusesoftclippedbases", default = false)
+      stand_call_conf = config("stand_call_conf", default = 5)
+      stand_emit_conf = config("stand_emit_conf", default = 0)
+    }
+    if (bamOutput != null && nct.getOrElse(1) > 1) {
+      nct = Option(1)
+      logger.warn("BamOutput is on, nct/threads is forced to set on 1, this option is only for debug")
+    }
+  }
+
+  def useGvcf() {
+    emitRefConfidence = org.broadinstitute.gatk.tools.walkers.haplotypecaller.ReferenceConfidenceMode.GVCF
+    variant_index_type = GATKVCFIndexType.LINEAR
+    variant_index_parameter = config("variant_index_parameter", default = 128000)
+  }
+}

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/IndelRealigner.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class IndelRealigner(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.IndelRealigner with GatkGeneral {
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+}
+
+object IndelRealigner {
+  def apply(root: Configurable, input: File, targetIntervals: File, outputDir: String): IndelRealigner = {
+    val ir = new IndelRealigner(root)
+    ir.input_file :+= input
+    ir.targetIntervals = targetIntervals
+    ir.out = new File(outputDir, input.getName.stripSuffix(".bam") + ".realign.bam")
+    return ir
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/PrintReads.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class PrintReads(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.PrintReads with GatkGeneral {
+  memoryLimit = Option(4)
+
+  override val defaultVmem = "8G"
+
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+}
+
+object PrintReads {
+  def apply(root: Configurable, input: File, output: File): PrintReads = {
+    val br = new PrintReads(root)
+    br.input_file :+= input
+    br.out = output
+    return br
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/RealignerTargetCreator.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class RealignerTargetCreator(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.RealignerTargetCreator with GatkGeneral {
+  override val defaultVmem = "6G"
+  memoryLimit = Some(2.5)
+
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+
+  if (config.contains("known")) known ++= config("known").asFileList
+}
+
+object RealignerTargetCreator {
+  def apply(root: Configurable, input: File, outputDir: String): RealignerTargetCreator = {
+    val re = new RealignerTargetCreator(root)
+    re.input_file :+= input
+    re.out = new File(outputDir, input.getName.stripSuffix(".bam") + ".realign.intervals")
+    return re
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/SelectVariants.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class SelectVariants(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.SelectVariants with GatkGeneral {
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+}
+
+object SelectVariants {
+  def apply(root: Configurable, input: File, output: File): SelectVariants = {
+    val sv = new SelectVariants(root)
+    sv.variant = input
+    sv.out = output
+    return sv
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/UnifiedGenotyper.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class UnifiedGenotyper(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.UnifiedGenotyper with GatkGeneral {
+  override def afterGraph {
+    super.afterGraph
+
+    genotype_likelihoods_model = org.broadinstitute.gatk.tools.walkers.genotyper.GenotypeLikelihoodsCalculationModel.Model.BOTH
+    if (config.contains("scattercount")) scatterCount = config("scattercount")
+    if (config.contains("dbsnp")) this.dbsnp = config("dbsnp")
+    this.sample_ploidy = config("ploidy")
+    nct = config("threads", default = 1)
+    memoryLimit = Option(nct.getOrElse(1) * 2)
+    if (config.contains("allSitePLs")) this.allSitePLs = config("allSitePLs")
+    if (config.contains("output_mode")) {
+      import org.broadinstitute.gatk.tools.walkers.genotyper.OutputMode._
+      config("output_mode").asString match {
+        case "EMIT_ALL_CONFIDENT_SITES" => output_mode = EMIT_ALL_CONFIDENT_SITES
+        case "EMIT_ALL_SITES"           => output_mode = EMIT_ALL_SITES
+        case "EMIT_VARIANTS_ONLY"       => output_mode = EMIT_VARIANTS_ONLY
+        case e                          => logger.warn("output mode '" + e + "' does not exist")
+      }
+    }
+
+    if (config("inputtype", default = "dna").asString == "rna") {
+      stand_call_conf = config("stand_call_conf", default = 5)
+      stand_emit_conf = config("stand_emit_conf", default = 0)
+    } else {
+      stand_call_conf = config("stand_call_conf", default = 5)
+      stand_emit_conf = config("stand_emit_conf", default = 0)
+    }
+  }
+}

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/VariantAnnotator.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class VariantAnnotator(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.VariantAnnotator with GatkGeneral {
+  if (config.contains("scattercount")) scatterCount = config("scattercount")
+  dbsnp = config("dbsnp")
+}
+
+object VariantAnnotator {
+  def apply(root: Configurable, input: File, bamFiles: List[File], output: File): VariantAnnotator = {
+    val va = new VariantAnnotator(root)
+    va.variant = input
+    va.input_file = bamFiles
+    va.out = output
+    return va
+  }
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/VariantEval.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+
+class VariantEval(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.VariantEval with GatkGeneral {
+  override def afterGraph {
+    super.afterGraph
+  }
+}
+
+object VariantEval {
+  def apply(root: Configurable, sample: File, compareWith: File,
+            output: File): VariantEval = {
+    val vareval = new VariantEval(root)
+    vareval.eval = Seq(sample)
+    vareval.comp = Seq(compareWith)
+    vareval.out = output
+    vareval.afterGraph
+    return vareval
+  }
+
+  def apply(root: Configurable, sample: File, compareWith: File,
+            output: File, ST: Seq[String], EV: Seq[String]): VariantEval = {
+    val vareval = new VariantEval(root)
+    vareval.eval = Seq(sample)
+    vareval.comp = Seq(compareWith)
+    vareval.out = output
+    vareval.noST = true
+    vareval.ST = ST
+    vareval.noEV = true
+    vareval.EV = EV
+    vareval.afterGraph
+    return vareval
+  }
+
+}
\ No newline at end of file

protected/biopet-gatk-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/gatk/VariantRecalibrator.scala

+package nl.lumc.sasc.biopet.extensions.gatk
+
+import java.io.File
+import nl.lumc.sasc.biopet.core.config.Configurable
+import org.broadinstitute.gatk.queue.extensions.gatk.TaggedFile
+
+class VariantRecalibrator(val root: Configurable) extends org.broadinstitute.gatk.queue.extensions.gatk.VariantRecalibrator with GatkGeneral {
+  nt = Option(getThreads(4))
+  memoryLimit = Option(nt.getOrElse(1) * 2)
+
+  if (config.contains("dbsnp")) resource :+= new TaggedFile(config("dbsnp").asString, "known=true,training=false,truth=false,prior=2.0")
+
+  an = config("annotation", default = List("QD", "DP", "FS", "ReadPosRankSum", "MQRankSum")).asStringList
+  minNumBadVariants = config("minnumbadvariants")
+  maxGaussians = config("maxgaussians")
+}
+
+object VariantRecalibrator {
+  def apply(root: Configurable, input: File, recal_file: File, tranches_file: File, indel: Boolean = false): VariantRecalibrator = {
+    val vr = new VariantRecalibrator(root) {
+      override lazy val configName = "variantrecalibrator"
+      override def configPath: List[String] = (if (indel) "indel" else "snp") :: super.configPath
+      if (indel) {
+        mode = org.broadinstitute.gatk.tools.walkers.variantrecalibration.VariantRecalibratorArgumentCollection.Mode.INDEL
+        defaults ++= Map("ts_filter_level" -> 99.0)
+        if (config.contains("mills")) resource :+= new TaggedFile(config("mills").asString, "known=false,training=true,truth=true,prior=12.0")
+      } else {
+        mode = org.broadinstitute.gatk.tools.walkers.variantrecalibration.VariantRecalibratorArgumentCollection.Mode.SNP
+        defaults ++= Map("ts_filter_level" -> 99.5)
+        if (config.contains("hapmap")) resource +:= new TaggedFile(config("hapmap").asString, "known=false,training=true,truth=true,prior=15.0")
+        if (config.contains("omni")) resource +:= new TaggedFile(config("omni").asString, "known=false,training=true,truth=true,prior=12.0")
+        if (config.contains("1000G")) resource +:= new TaggedFile(config("1000G").asString, "known=false,training=true,truth=false,prior=10.0")
+      }
+    }
+    vr.input :+= input
+    vr.recal_file = recal_file
+    vr.tranches_file = tranches_file
+    return vr
+  }
+}
\ No newline at end of file

protected/pom.xml

@@ -8,6 +8,9 @@
     </parent>
     <artifactId>protected</artifactId>
     <packaging>pom</packaging>
+    <modules>
+        <module>biopet-gatk-extensions</module>
+    </modules>
     <properties>
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
     </properties>