First wrappers for kopisu #106

public/biopet-framework/src/main/scala/nl/lumc/sasc/biopet/extensions/conifer/Conifer.scala

+/**
+ * Biopet is built on top of GATK Queue for building bioinformatic
+ * pipelines. It is mainly intended to support LUMC SHARK cluster which is running
+ * SGE. But other types of HPC that are supported by GATK Queue (such as PBS)
+ * should also be able to execute Biopet tools and pipelines.
+ *
+ * Copyright 2014 Sequencing Analysis Support Core - Leiden University Medical Center
+ *
+ * Contact us at: sasc@lumc.nl
+ *
+ * A dual licensing mode is applied. The source code within this project that are
+ * not part of GATK Queue is freely available for non-commercial use under an AGPL
+ * license; For commercial users or users who do not want to follow the AGPL
+ * license, please contact us to obtain a separate license.
+ */
+package nl.lumc.sasc.biopet.extensions.conifer
+
+import nl.lumc.sasc.biopet.core.BiopetCommandLineFunction
+
+abstract class Conifer extends BiopetCommandLineFunction {
+  override def subPath = "conifer" :: super.subPath
+  executable = config("exe", default = "conifer")
+  override val versionRegex = """(.*)""".r
+  override val versionExitcode = List(0)
+  override def versionCommand = executable + " --version"
+}

public/biopet-framework/src/main/scala/nl/lumc/sasc/biopet/extensions/conifer/ConiferAnalyze.scala

+/**
+ * Biopet is built on top of GATK Queue for building bioinformatic
+ * pipelines. It is mainly intended to support LUMC SHARK cluster which is running
+ * SGE. But other types of HPC that are supported by GATK Queue (such as PBS)
+ * should also be able to execute Biopet tools and pipelines.
+ *
+ * Copyright 2014 Sequencing Analysis Support Core - Leiden University Medical Center
+ *
+ * Contact us at: sasc@lumc.nl
+ *
+ * A dual licensing mode is applied. The source code within this project that are
+ * not part of GATK Queue is freely available for non-commercial use under an AGPL
+ * license; For commercial users or users who do not want to follow the AGPL
+ * license, please contact us to obtain a separate license.
+ */
+package nl.lumc.sasc.biopet.extensions.conifer
+
+import java.io.File
+
+import nl.lumc.sasc.biopet.core.config.Configurable
+import org.broadinstitute.gatk.utils.commandline.{Argument, Input, Output}
+
+class ConiferAnalyze(val root: Configurable) extends Conifer {
+
+  @Input(doc = "Probes / capture kit definition as bed file: chr,start,stop,gene-annot", required = true)
+  var probes: File = _
+
+  @Input(doc = "Path to Conifer RPKM files", required = true)
+  var rpkm_dir: File = _
+
+  @Output(doc = "Output RPKM.txt", shortName = "out")
+  var output: File = _
+
+  @Argument(doc = "SVD, number of components to remove", minRecommendedValue = 2, maxRecommendedValue = 5,
+    minValue = 2, maxValue = 20)
+  var svd: Option[Int] = config("svd")
+
+  @Argument(doc="Minimum population median RPKM per probe")
+  var min_rpkm: Option[Double] = config("min_rpkm")
+
+  override def afterGraph {
+    this.checkExecutable
+  }
+
+  def cmdLine = required(executable) +
+    required("rpkm")+
+    " --probes" + required(probes) +
+    " --rpkm_dir" + required(rpkm_dir) +
+    " --output" + required(output) +
+    optional("--svd",svd) +
+    optional("--min_rpkm", min_rpkm)
+}

public/biopet-framework/src/main/scala/nl/lumc/sasc/biopet/extensions/conifer/ConiferCall.scala

+/**
+ * Biopet is built on top of GATK Queue for building bioinformatic
+ * pipelines. It is mainly intended to support LUMC SHARK cluster which is running
+ * SGE. But other types of HPC that are supported by GATK Queue (such as PBS)
+ * should also be able to execute Biopet tools and pipelines.
+ *
+ * Copyright 2014 Sequencing Analysis Support Core - Leiden University Medical Center
+ *
+ * Contact us at: sasc@lumc.nl
+ *
+ * A dual licensing mode is applied. The source code within this project that are
+ * not part of GATK Queue is freely available for non-commercial use under an AGPL
+ * license; For commercial users or users who do not want to follow the AGPL
+ * license, please contact us to obtain a separate license.
+ */
+package nl.lumc.sasc.biopet.extensions.conifer
+
+import java.io.File
+
+import nl.lumc.sasc.biopet.core.config.Configurable
+import org.broadinstitute.gatk.utils.commandline.{Argument, Input, Output}
+
+class ConiferCall(val root: Configurable) extends Conifer {
+
+  @Input(doc = "Input analysis.hdf5", required = true)
+  var input: File = _
+
+  @Output(doc = "Output calls.txt", shortName = "out")
+  var output: File = _
+
+  override def afterGraph {
+    this.checkExecutable
+  }
+
+  def cmdLine = required(executable) +
+    required("call")+
+    " --input" + required(input) +
+    " --output" + required(output)
+}

public/biopet-framework/src/main/scala/nl/lumc/sasc/biopet/extensions/conifer/ConiferRPKM.scala

+/**
+ * Biopet is built on top of GATK Queue for building bioinformatic
+ * pipelines. It is mainly intended to support LUMC SHARK cluster which is running
+ * SGE. But other types of HPC that are supported by GATK Queue (such as PBS)
+ * should also be able to execute Biopet tools and pipelines.
+ *
+ * Copyright 2014 Sequencing Analysis Support Core - Leiden University Medical Center
+ *
+ * Contact us at: sasc@lumc.nl
+ *
+ * A dual licensing mode is applied. The source code within this project that are
+ * not part of GATK Queue is freely available for non-commercial use under an AGPL
+ * license; For commercial users or users who do not want to follow the AGPL
+ * license, please contact us to obtain a separate license.
+ */
+package nl.lumc.sasc.biopet.extensions.conifer
+
+import java.io.File
+
+import nl.lumc.sasc.biopet.core.config.Configurable
+import org.broadinstitute.gatk.utils.commandline.{Output, Input}
+
+class ConiferRPKM(val root: Configurable) extends Conifer {
+
+  @Input(doc = "Bam file", required = true)
+  var bamFile: File = _
+
+  @Input(doc = "Probes / capture kit definition as bed file: chr,start,stop,gene-annot", required = true)
+  var probes: File = _
+
+  @Output(doc = "Output RPKM.txt", shortName = "out")
+  var output: File = _
+
+  private var config_file: File = _
+
+  override def afterGraph {
+    this.checkExecutable
+  }
+
+  def cmdLine = required(executable) +
+    required("rpkm")+
+    " --probes" + required(probes) +
+    " --input" + required(bamFile) +
+    " --output" + required(output)
+}

public/kopisu/.gitignore

+/target/
\ No newline at end of file

public/kopisu/Kopisu.iml

+<?xml version="1.0" encoding="UTF-8"?>
+<module org.jetbrains.idea.maven.project.MavenProjectsManager.isMavenModule="true" type="JAVA_MODULE" version="4">
+  <component name="NewModuleRootManager" inherit-compiler-output="false">
+    <output url="file://$MODULE_DIR$/target/classes" />
+    <output-test url="file://$MODULE_DIR$/target/test-classes" />
+    <content url="file://$MODULE_DIR$">
+      <sourceFolder url="file://$MODULE_DIR$/src/main/resources" type="java-resource" />
+      <sourceFolder url="file://$MODULE_DIR$/src/main/scala" isTestSource="false" />
+      <excludeFolder url="file://$MODULE_DIR$/target" />
+    </content>
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+    <orderEntry type="module" module-name="BiopetFramework" />
+    <orderEntry type="library" name="Maven: org.scala-lang:scala-library:2.11.2" level="project" />
+    <orderEntry type="library" name="Maven: org.broadinstitute.gatk:gatk-queue:3.3" level="project" />
+    <orderEntry type="library" name="Maven: org.broadinstitute.gatk:gatk-tools-public:3.3" level="project" />
+    <orderEntry type="library" name="Maven: org.broadinstitute.gatk:gatk-engine:3.3" level="project" />
+    <orderEntry type="library" name="Maven: org.scala-lang:scala-compiler:2.10.2" level="project" />
+    <orderEntry type="library" name="Maven: org.scala-lang:scala-reflect:2.10.2" level="project" />
+    <orderEntry type="library" name="Maven: log4j:log4j:1.2.15" level="project" />
+    <orderEntry type="library" name="Maven: net.sf.jgrapht:jgrapht:0.8.3" level="project" />
+    <orderEntry type="library" name="Maven: org.apache.commons:commons-email:1.2" level="project" />
+    <orderEntry type="library" name="Maven: javax.activation:activation:1.1" level="project" />
+    <orderEntry type="library" name="Maven: javax.mail:mail:1.4.4" level="project" />
+    <orderEntry type="module-library">
+      <library name="Maven: com.sun:tools:1.4.2">
+        <CLASSES>
+          <root url="jar:///usr/lib/jvm/java-7-openjdk-amd64/lib/tools.jar!/" />
+        </CLASSES>
+        <JAVADOC />
+        <SOURCES />
+      </library>
+    </orderEntry>
+    <orderEntry type="library" name="Maven: org.broadinstitute.gatk:gatk-queue-extensions-distribution:3.3" level="project" />
+    <orderEntry type="library" name="Maven: org.broadinstitute.gatk:gatk-tools-protected:3.3" level="project" />
+    <orderEntry type="library" name="Maven: gov.nist.math:jama:1.0.2" level="project" />
+    <orderEntry type="library" scope="RUNTIME" name="Maven: org.broadinstitute.gatk:gatk-queue-extensions-generator:3.3" level="project" />
+    <orderEntry type="library" name="Maven: org.broadinstitute.gatk:gatk-utils:3.3" level="project" />
+    <orderEntry type="library" name="Maven: samtools:htsjdk:1.120.1620" level="project" />
+    <orderEntry type="library" name="Maven: org.xerial.snappy:snappy-java:1.0.3-rc3" level="project" />
+    <orderEntry type="library" name="Maven: picard:picard:1.120.1579" level="project" />
+    <orderEntry type="library" name="Maven: org.apache.ant:ant:1.8.2" level="project" />
+    <orderEntry type="module-library">
+      <library name="Maven: com.sun:tools.jar:1.5">
+        <CLASSES>
+          <root url="jar:///usr/lib/jvm/java-7-openjdk-amd64/lib/tools.jar!/" />
+        </CLASSES>
+        <JAVADOC />
+        <SOURCES />
+      </library>
+    </orderEntry>
+    <orderEntry type="library" name="Maven: colt:colt:1.2.0" level="project" />
+    <orderEntry type="library" name="Maven: concurrent:concurrent:1.3.4" level="project" />
+    <orderEntry type="library" name="Maven: it.unimi.dsi:fastutil:6.5.3" level="project" />
+    <orderEntry type="library" name="Maven: org.simpleframework:simple-xml:2.0.4" level="project" />
+    <orderEntry type="library" name="Maven: stax:stax-api:1.0.1" level="project" />
+    <orderEntry type="library" name="Maven: stax:stax:1.2.0" level="project" />
+    <orderEntry type="library" name="Maven: org.reflections:reflections:0.9.9-RC1" level="project" />
+    <orderEntry type="library" name="Maven: org.javassist:javassist:3.16.1-GA" level="project" />
+    <orderEntry type="library" name="Maven: dom4j:dom4j:1.6.1" level="project" />
+    <orderEntry type="library" name="Maven: xml-apis:xml-apis:1.0.b2" level="project" />
+    <orderEntry type="library" name="Maven: org.slf4j:slf4j-log4j12:1.6.1" level="project" />
+    <orderEntry type="library" name="Maven: org.slf4j:slf4j-api:1.6.1" level="project" />
+    <orderEntry type="library" name="Maven: org.freemarker:freemarker:2.3.18" level="project" />
+    <orderEntry type="library" name="Maven: org.apache.commons:commons-jexl:2.1.1" level="project" />
+    <orderEntry type="library" name="Maven: commons-logging:commons-logging:1.1.1" level="project" />
+    <orderEntry type="library" name="Maven: commons-lang:commons-lang:2.5" level="project" />
+    <orderEntry type="library" name="Maven: commons-io:commons-io:2.1" level="project" />
+    <orderEntry type="library" name="Maven: commons-collections:commons-collections:3.2.1" level="project" />
+    <orderEntry type="library" name="Maven: org.apache.commons:commons-math:2.2" level="project" />
+    <orderEntry type="library" name="Maven: net.java.dev.jna:jna:3.2.7" level="project" />
+    <orderEntry type="library" name="Maven: net.java.dev.jets3t:jets3t:0.8.1" level="project" />
+    <orderEntry type="library" name="Maven: commons-codec:commons-codec:1.3" level="project" />
+    <orderEntry type="library" name="Maven: commons-httpclient:commons-httpclient:3.1" level="project" />
+    <orderEntry type="library" name="Maven: com.jamesmurty.utils:java-xmlbuilder:0.4" level="project" />
+    <orderEntry type="library" name="Maven: us.levk:drmaa-gridengine:6.2u5" level="project" />
+    <orderEntry type="library" name="Maven: us.levk:drmaa-common:1.0" level="project" />
+    <orderEntry type="library" name="Maven: com.google.code.gson:gson:2.2.2" level="project" />
+    <orderEntry type="library" name="Maven: org.apache.httpcomponents:httpclient:4.1.1" level="project" />
+    <orderEntry type="library" name="Maven: org.apache.httpcomponents:httpcore:4.1" level="project" />
+    <orderEntry type="library" name="Maven: com.google.code.cofoja:cofoja:1.0-r139" level="project" />
+    <orderEntry type="library" name="Maven: io.argonaut:argonaut_2.11:6.1-M4" level="project" />
+    <orderEntry type="library" name="Maven: org.scalaz:scalaz-core_2.11:7.1.0" level="project" />
+    <orderEntry type="library" name="Maven: org.scala-lang.modules:scala-parser-combinators_2.11:1.0.2" level="project" />
+    <orderEntry type="library" name="Maven: org.scala-lang.modules:scala-xml_2.11:1.0.2" level="project" />
+    <orderEntry type="library" name="Maven: com.github.julien-truffaut:monocle-core_2.11:0.5.0" level="project" />
+    <orderEntry type="library" name="Maven: org.biojava:biojava3-core:3.1.0" level="project" />
+    <orderEntry type="library" name="Maven: org.biojava:biojava3-sequencing:3.1.0" level="project" />
+    <orderEntry type="library" name="Maven: com.google.guava:guava:18.0" level="project" />
+    <orderEntry type="library" name="Maven: com.github.scopt:scopt_2.10:3.2.0" level="project" />
+  </component>
+</module>
\ No newline at end of file

public/kopisu/pom.xml

+<!--
+
+    Biopet is built on top of GATK Queue for building bioinformatic
+    pipelines. It is mainly intended to support LUMC SHARK cluster which is running
+    SGE. But other types of HPC that are supported by GATK Queue (such as PBS)
+    should also be able to execute Biopet tools and pipelines.
+
+    Copyright 2014 Sequencing Analysis Support Core - Leiden University Medical Center
+
+    Contact us at: sasc@lumc.nl
+
+    A dual licensing mode is applied. The source code within this project that are
+    not part of GATK Queue is freely available for non-commercial use under an AGPL
+    license; For commercial users or users who do not want to follow the AGPL
+    license, please contact us to obtain a separate license.
+
+-->
+<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>Kopisu</artifactId>
+    <packaging>jar</packaging>
+    
+    <parent>
+        <groupId>nl.lumc.sasc</groupId>
+        <artifactId>Biopet</artifactId>
+        <version>0.3.0-DEV</version>
+        <relativePath>../</relativePath>
+    </parent>
+    
+    <inceptionYear>2014</inceptionYear>
+    <name>Kopisu</name>
+
+    <dependencies>
+        <dependency>
+            <groupId>nl.lumc.sasc</groupId>
+            <artifactId>BiopetFramework</artifactId>
+            <version>${project.version}</version>
+        </dependency>
+    </dependencies>
+</project>

public/kopisu/src/main/scala/nl/lumc/sasc/biopet/pipelines/kopisu/Kopisu.scala

+/**
+ * Biopet is built on top of GATK Queue for building bioinformatic
+ * pipelines. It is mainly intended to support LUMC SHARK cluster which is running
+ * SGE. But other types of HPC that are supported by GATK Queue (such as PBS)
+ * should also be able to execute Biopet tools and pipelines.
+ *
+ * Copyright 2014 Sequencing Analysis Support Core - Leiden University Medical Center
+ *
+ * Contact us at: sasc@lumc.nl
+ *
+ * A dual licensing mode is applied. The source code within this project that are
+ * not part of GATK Queue is freely available for non-commercial use under an AGPL
+ * license; For commercial users or users who do not want to follow the AGPL
+ * license, please contact us to obtain a separate license.
+ */
+package nl.lumc.sasc.biopet.pipelines.sage
+
+import nl.lumc.sasc.biopet.core.{ MultiSampleQScript, PipelineCommand }
+import nl.lumc.sasc.biopet.core.config.Configurable
+import nl.lumc.sasc.biopet.extensions.Cat
+import nl.lumc.sasc.biopet.extensions.bedtools.BedtoolsCoverage
+import nl.lumc.sasc.biopet.extensions.picard.MergeSamFiles
+import nl.lumc.sasc.biopet.pipelines.flexiprep.Flexiprep
+import nl.lumc.sasc.biopet.pipelines.mapping.Mapping
+import nl.lumc.sasc.biopet.tools.PrefixFastq
+import nl.lumc.sasc.biopet.tools.BedtoolsCoverageToCounts
+import nl.lumc.sasc.biopet.scripts.SquishBed
+import nl.lumc.sasc.biopet.tools.SageCountFastq
+import nl.lumc.sasc.biopet.tools.SageCreateLibrary
+import nl.lumc.sasc.biopet.tools.SageCreateTagCounts
+import org.broadinstitute.gatk.queue.QScript
+
+class Kopisu(val root: Configurable) extends QScript with MultiSampleQScript {
+  def this() = this(null)
+
+  @Input(doc = "Input bamfile", required = true)
+  var bamFile: File = config("bam")
+
+  class LibraryOutput extends AbstractLibraryOutput {
+  }
+
+  class SampleOutput extends AbstractSampleOutput {
+  }
+
+  def init() {
+    if (!outputDir.endsWith("/")) outputDir += "/"
+  }
+
+  def biopetScript() {
+    runSamplesJobs
+  }
+
+  // Called for each sample
+  def runSingleSampleJobs(sampleConfig: Map[String, Any]): SampleOutput = {
+    val sampleOutput = new SampleOutput
+    var libraryBamfiles: List[File] = List()
+    var libraryFastqFiles: List[File] = List()
+    val sampleID: String = sampleConfig("ID").toString
+    val sampleDir: String = globalSampleDir + sampleID + "/"
+
+    for ((library, libraryFiles) <- runLibraryJobs(sampleConfig)) {
+    }
+
+    return sampleOutput
+  }
+
+  // Called for each run from a sample
+  def runSingleLibraryJobs(runConfig: Map[String, Any], sampleConfig: Map[String, Any]): LibraryOutput = {
+    val libraryOutput = new LibraryOutput
+    val runID: String = runConfig("ID").toString
+    val sampleID: String = sampleConfig("ID").toString
+    val runDir: String = globalSampleDir + sampleID + "/run_" + runID + "/"
+    if (runConfig.contains("bam")) {
+    } else this.logger.error("Sample: " + sampleID + ": No R1 found for run: " + runConfig)
+    return libraryOutput
+  }
+}
+
+object Kopisu extends PipelineCommand