Property Detail

• static final int MINIMUM_TARGET_REGIONS

• graxxia.Matrix covs

• gngs.Region debugRegion

• double deletionCallThreshold

  Likelihood ratio at which to output deletions

• double deletionRate

  An indicator of the number deletion calls expected per target region (or, approximately, the probability any given target is deleted state, though that interpretation does not account for multi-target calls)

• double duplicationCallThreshold

  Likelihood ratio at which to output duplications

• int maxPCAComponents

  The maximum number of PCA components that will be removed

• int minExpectedDeletionCallsPerChromosome

  The expected number of deletion calls is calculated from the per-target region rate but this can lead to an unreasonably low bar for small chromosomes

• java.io.Writer optimisationWriter

• java.util.Map<java.lang.String, Regions> results

• double[] targetRegionGCRfractions

• Regions targets

Constructor Detail

• Delfin()

Method Detail

• final Regions analyseChromosome(java.lang.String chr)

• void calculateIntervalGCs()

• java.util.List<java.util.List<java.lang.Double>> callAndOptimise(java.lang.String chr, java.lang.String sample, graxxia.Matrix scaled, graxxia.Matrix std, graxxia.Matrix reduced)

  Tries calling deletions while removing increasing numbers of prinicipal components. When an acceptably small number is returned, stops and returns both duplications and deletions.

  Returns:

  List containing two elements: [indexes of called deletions, indexes of called duplications]

• @groovy.transform.CompileStatic void initGCNormalisation()

• graxxia.Matrix loadCoverage(java.io.File path)

  Loads per-target-region coverage depth from the given provided files, which are in the form of an initial column containing the sample and then a column for each target region containing the read depth over that exon.

  Parameters:

  path - path to the bed file

  Returns:

  a Matrix with a row per sample, normalised so that the mean of each row is 1

• static void main(java.lang.Object args)

• @groovy.transform.CompileStatic java.util.List<Matrix> normaliseAndComputeLikelihoods(java.lang.String sample, java.lang.String chr, graxxia.Matrix scaled, graxxia.Matrix std, graxxia.Matrix reduced, int nComponentsToRemove)

  Transform the matrix of row-normalised coverage values into a matrix of log-scaled likelihood ratios comparing the z-score of a diploid model to a single-ploidy model.

  Parameters:

  row-normalised - coverage values, row per sample, column per target region

  Returns:

  matrix of likelihoods

• @java.lang.Override void run()