Package picard.sam.markduplicates

Class EstimateLibraryComplexity

java.lang.Object
picard.cmdline.CommandLineProgram
picard.sam.markduplicates.util.AbstractOpticalDuplicateFinderCommandLineProgram
picard.sam.markduplicates.EstimateLibraryComplexity
@DocumentedFeature public class EstimateLibraryComplexity extends AbstractOpticalDuplicateFinderCommandLineProgram

Attempts to estimate library complexity from sequence alone. Does so by sorting all reads by the first N bases (5 by default) of each read and then comparing reads with the first N bases identical to each other for duplicates. Reads are considered to be duplicates if they match each other with no gaps and an overall mismatch rate less than or equal to MAX_DIFF_RATE (0.03 by default).

Reads of poor quality are filtered out so as to provide a more accurate estimate. The filtering removes reads with any no-calls in the first N bases or with a mean base quality lower than MIN_MEAN_QUALITY across either the first or second read.

The algorithm attempts to detect optical duplicates separately from PCR duplicates and excludes these in the calculation of library size. Also, since there is no alignment to screen out technical reads one further filter is applied on the data. After examining all reads a Histogram is built of [#reads in duplicate set -> #of duplicate sets]; all bins that contain exactly one duplicate set are then removed from the Histogram as outliers before library size is estimated.

  • Field Details

    • INPUT

      @Argument(shortName="I", doc="One or more files to combine and estimate library complexity from. Reads can be mapped or unmapped.") public List<File> INPUT

    • OUTPUT

      @Argument(shortName="O", doc="Output file to writes per-library metrics to.") public File OUTPUT

    • MIN_IDENTICAL_BASES

      @Argument(doc="The minimum number of bases at the starts of reads that must be identical for reads to be grouped together for duplicate detection. In effect total_reads / 4^max_id_bases reads will be compared at a time, so lower numbers will produce more accurate results but consume exponentially more memory and CPU.") public int MIN_IDENTICAL_BASES

    • MAX_DIFF_RATE

      @Argument(doc="The maximum rate of differences between two reads to call them identical.") public double MAX_DIFF_RATE

    • MIN_MEAN_QUALITY

      @Argument(doc="The minimum mean quality of the bases in a read pair for the read to be analyzed. Reads with lower average quality are filtered out and not considered in any calculations.") public int MIN_MEAN_QUALITY

    • MAX_GROUP_RATIO

      @Argument(doc="Do not process self-similar groups that are this many times over the mean expected group size. I.e. if the input contains 10m read pairs and MIN_IDENTICAL_BASES is set to 5, then the mean expected group size would be approximately 10 reads.") public int MAX_GROUP_RATIO

    • BARCODE_TAG

      @Argument(doc="Barcode SAM tag (ex. BC for 10X Genomics)", optional=true) public String BARCODE_TAG

    • READ_ONE_BARCODE_TAG

      @Argument(doc="Read one barcode SAM tag (ex. BX for 10X Genomics)", optional=true) public String READ_ONE_BARCODE_TAG

    • READ_TWO_BARCODE_TAG

      @Argument(doc="Read two barcode SAM tag (ex. BX for 10X Genomics)", optional=true) public String READ_TWO_BARCODE_TAG

    • MAX_READ_LENGTH

      @Argument(doc="The maximum number of bases to consider when comparing reads (0 means no maximum).", optional=true) public int MAX_READ_LENGTH

    • MIN_GROUP_COUNT

      @Argument(doc="Minimum number group count. On a per-library basis, we count the number of groups of duplicates that have a particular size. Omit from consideration any count that is less than this value. For example, if we see only one group of duplicates with size 500, we omit it from the metric calculations if MIN_GROUP_COUNT is set to two. Setting this to two may help remove technical artifacts from the library size calculation, for example, adapter dimers.", optional=true) public int MIN_GROUP_COUNT

  • Constructor Details

    • EstimateLibraryComplexity

      public EstimateLibraryComplexity()

  • Method Details

    • customCommandLineValidation

      protected String[] customCommandLineValidation()
      Description copied from class: CommandLineProgram
      Put any custom command-line validation in an override of this method. clp is initialized at this point and can be used to print usage and access argv. Any options set by command-line parser can be validated.
      Overrides:
      customCommandLineValidation in class AbstractOpticalDuplicateFinderCommandLineProgram
      Returns:
      null if command line is valid. If command line is invalid, returns an array of error message to be written to the appropriate place.

    • getBarcodeValue

      public int getBarcodeValue(htsjdk.samtools.SAMRecord record)

    • getReadBarcodeValue

      public static int getReadBarcodeValue(htsjdk.samtools.SAMRecord record, String tag)

    • doWork

      protected int doWork()
      Method that does most of the work. Reads through the input BAM file and extracts the read sequences of each read pair and sorts them via a SortingCollection. Then traverses the sorted reads and looks at small groups at a time to find duplicates.
      Specified by:
      doWork in class CommandLineProgram
      Returns:
      program exit status.