org.broadinstitute.hellbender.engine.spark

Class LocusWalkerSpark

java.lang.Object

org.broadinstitute.hellbender.cmdline.CommandLineProgram

org.broadinstitute.hellbender.engine.spark.SparkCommandLineProgram

org.broadinstitute.hellbender.engine.spark.GATKSparkTool

org.broadinstitute.hellbender.engine.spark.LocusWalkerSpark

All Implemented Interfaces:

java.io.Serializable, org.broadinstitute.barclay.argparser.CommandLinePluginProvider

Direct Known Subclasses:

CollectAllelicCountsSpark, ExampleLocusWalkerSpark, PileupSpark

public abstract class LocusWalkerSpark
extends GATKSparkTool

A Spark version of LocusWalker. Subclasses should implement processAlignments(JavaRDD, JavaSparkContext) and operate on the passed in RDD.

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from class org.broadinstitute.hellbender.engine.spark.GATKSparkTool

GATKSparkTool.ReadInputMergingPolicy

Field Summary

Fields

Modifier and Type	Field and Description
protected int	maxDepthPerSample
int	readShardSize
boolean	shuffle

Fields inherited from class org.broadinstitute.hellbender.engine.spark.GATKSparkTool

addOutputVCFCommandLine, BAM_PARTITION_SIZE_LONG_NAME, bamPartitionSplitSize, CREATE_OUTPUT_BAM_SPLITTING_INDEX_LONG_NAME, createOutputBamIndex, createOutputBamSplittingIndex, createOutputVariantIndex, features, intervalArgumentCollection, NUM_REDUCERS_LONG_NAME, numReducers, OUTPUT_SHARD_DIR_LONG_NAME, readArguments, referenceArguments, sequenceDictionaryValidationArguments, SHARDED_OUTPUT_LONG_NAME, shardedOutput, shardedPartsDir, USE_NIO, useNio

Fields inherited from class org.broadinstitute.hellbender.engine.spark.SparkCommandLineProgram

programName, SPARK_PROGRAM_NAME_LONG_NAME, sparkArgs

Fields inherited from class org.broadinstitute.hellbender.cmdline.CommandLineProgram

GATK_CONFIG_FILE, logger, NIO_MAX_REOPENS, NIO_PROJECT_FOR_REQUESTER_PAYS, QUIET, specialArgumentsCollection, tmpDir, useJdkDeflater, useJdkInflater, VERBOSITY

Constructor Summary

Constructors

Constructor and Description
LocusWalkerSpark()

Method Summary

Modifier and Type	Method and Description
protected int	defaultMaxDepthPerSample() Returns default value for the maxDepthPerSample parameter, if none is provided on the command line.
boolean	emitEmptyLoci() Does this tool emit information for uncovered loci? Tools that do should override to return true.
org.apache.spark.api.java.JavaRDD<LocusWalkerContext>	getAlignments(org.apache.spark.api.java.JavaSparkContext ctx) Loads alignments and the corresponding reference and features into a JavaRDD for the intervals specified.
protected LIBSDownsamplingInfo	getDownsamplingInfo() Returns the downsampling info using maxDepthPerSample as target coverage.
protected abstract void	processAlignments(org.apache.spark.api.java.JavaRDD<LocusWalkerContext> rdd, org.apache.spark.api.java.JavaSparkContext ctx) Process the alignments and write output.
boolean	requiresReads() Does this tool require reads? Tools that do should override to return true.
protected void	runTool(org.apache.spark.api.java.JavaSparkContext ctx) Runs the tool itself after initializing and validating inputs.

Methods inherited from class org.broadinstitute.hellbender.engine.spark.GATKSparkTool

addReferenceFilesForSpark, addVCFsForSpark, editIntervals, getBestAvailableSequenceDictionary, getDefaultReadFilters, getDefaultToolVCFHeaderLines, getDefaultVariantAnnotationGroups, getDefaultVariantAnnotations, getGatkReadJavaRDD, getHeaderForReads, getIntervals, getPluginDescriptors, getReadInputMergingPolicy, getReads, getReadSourceHeaderMap, getReadSourceName, getRecommendedNumReducers, getReference, getReferenceSequenceDictionary, getReferenceWindowFunction, getSequenceDictionaryValidationArgumentCollection, getTargetPartitionSize, getUnfilteredReads, hasReads, hasReference, hasUserSuppliedIntervals, makeReadFilter, makeReadFilter, makeVariantAnnotations, requiresIntervals, requiresReference, runPipeline, useVariantAnnotations, validateSequenceDictionaries, writeReads, writeReads

Methods inherited from class org.broadinstitute.hellbender.engine.spark.SparkCommandLineProgram

afterPipeline, doWork, getProgramName

Methods inherited from class org.broadinstitute.hellbender.cmdline.CommandLineProgram

customCommandLineValidation, getCommandLine, getCommandLineParser, getDefaultHeaders, getMetricsFile, getSupportInformation, getToolkitName, getToolkitShortName, getToolStatusWarning, getUsage, getVersion, instanceMain, instanceMainPostParseArgs, isBetaFeature, isExperimentalFeature, onShutdown, onStartup, parseArgs, printLibraryVersions, printSettings, printStartupMessage, runTool, setDefaultHeaders, warnOnToolStatus

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

maxDepthPerSample

@Argument(fullName="max-depth-per-sample",
          shortName="max-depth-per-sample",
          doc="Maximum number of reads to retain per sample per locus. Reads above this threshold will be downsampled. Set to 0 to disable.",
          optional=true)
protected int maxDepthPerSample

readShardSize

@Argument(fullName="read-shard-size",
          shortName="read-shard-size",
          doc="Maximum size of each read shard, in bases.",
          optional=true)
public int readShardSize

shuffle

@Argument(doc="whether to use the shuffle implementation or overlaps partitioning (the default)",
          shortName="shuffle",
          fullName="shuffle",
          optional=true)
public boolean shuffle

Constructor Detail

LocusWalkerSpark

public LocusWalkerSpark()

Method Detail

defaultMaxDepthPerSample

protected int defaultMaxDepthPerSample()

Returns default value for the maxDepthPerSample parameter, if none is provided on the command line. Default implementation returns 0 (no downsampling by default).

requiresReads

public boolean requiresReads()

Description copied from class: GATKSparkTool

Does this tool require reads? Tools that do should override to return true.

Overrides:

requiresReads in class GATKSparkTool

Returns:

true if this tool requires reads, otherwise false

getDownsamplingInfo

protected final LIBSDownsamplingInfo getDownsamplingInfo()

Returns the downsampling info using maxDepthPerSample as target coverage.

emitEmptyLoci

public boolean emitEmptyLoci()

Does this tool emit information for uncovered loci? Tools that do should override to return true. NOTE: Typically, this should only be used when intervals are specified. NOTE: If MappedReadFilter is removed, then emitting empty loci will fail. NOTE: If there is no available sequence dictionary and this is set to true, there should be a failure. Please consider requiring reads and/or references for all tools that wish to set this to true.

Returns:

true if this tool requires uncovered loci information to be emitted, false otherwise

getAlignments

public org.apache.spark.api.java.JavaRDD<LocusWalkerContext> getAlignments(org.apache.spark.api.java.JavaSparkContext ctx)

Loads alignments and the corresponding reference and features into a JavaRDD for the intervals specified. If no intervals were specified, returns all the alignments.

Returns:

all alignments as a JavaRDD, bounded by intervals if specified.

runTool

protected void runTool(org.apache.spark.api.java.JavaSparkContext ctx)

Description copied from class: GATKSparkTool

Runs the tool itself after initializing and validating inputs. Must be implemented by subclasses.

Specified by:

runTool in class GATKSparkTool

Parameters:

ctx - our Spark context

processAlignments

protected abstract void processAlignments(org.apache.spark.api.java.JavaRDD<LocusWalkerContext> rdd,
                                          org.apache.spark.api.java.JavaSparkContext ctx)

Process the alignments and write output. Must be implemented by subclasses.

Parameters:

rdd - a distributed collection of LocusWalkerContext

ctx - our Spark context