Package org.gorpipe.gor.gava

Class VariantAssociation

java.lang.Object
org.gorpipe.gor.gava.VariantAssociation
public class VariantAssociation extends Object
Variant association algorithm, based on the VAAST algorithm. http://www.ncbi.nlm.nih.gov/pubmed/21700766

Computes a p-value for a genomic feature (typically a single gene) using a randomization test, where the affection status of the subjects is permuted and the test statistic computed using a composite likelihood ratio test as described in the VAAST paper.

Note: This is not an exact implementation and for example we have a bailout parameter to avoid spending to much computing time on uninteresting features.

Usage: Need to first set the case and control lists, the disease model (regular, dominant, or recessive), and other parameters. Then the procedure is as follows: - Before processing a feature, call initializeGroup() to clear the counts. - Iterate through the data for the feature and for each variant found for some subject, add the call counts using addFeature(). - Calculate the p-value (and a few other values) using calculateValues(). Repeat the process for the next feature.

  • Field Details

    • includeProtective

      protected boolean includeProtective

  • Constructor Details

    • VariantAssociation

      public VariantAssociation()
      Instantiate the class.

  • Method Details

    • setPnLists

      public void setPnLists(String[] cases, String[] controls)
      Parameters:
      cases - the cases
      controls - the controls

    • setModel

      public void setModel(boolean recessiveDisease, boolean dominantDisease, int controlPenetrance, int casePenetrance, boolean noMaxAlleleCounts, double upperFreqThreshold, boolean includeProtective)
      Set the disease model (inheritance and penetrance patterns).
      Parameters:
      recessiveDisease - true if recessive disease
      dominantDisease - true if dominant disease
      controlPenetrance - maximum count among controls, -1 for no constraint
      casePenetrance - true if recessive/dominant with no locus heterogeneity
      noMaxAlleleCounts - true if no max allele count for each case in dominant/recessive models
      upperFreqThreshold - upper threshold for the allele frequency
      includeProtective - true to include protective variants, false to exclude

    • setNumRandomIterations

      public void setNumRandomIterations(int maxIterations, int bailOutAfter)
      Parameters:
      maxIterations - the maximum number of random permutations for the p-value
      bailOutAfter - number of hits to trigger bailout

    • setCollapseThreshold

      public void setCollapseThreshold(int collapseThreshold)
      Parameters:
      collapseThreshold - collapse variants with fewer than this number of alleles among affecteds

    • setCancelMonitor

      public void setCancelMonitor(org.gorpipe.gor.monitor.CancelMonitor cancelMonitor)
      Parameters:
      cancelMonitor - cancel monitor

    • setDebug

      public void setDebug(boolean debug)
      Parameters:
      debug - if true, print out debug information

    • initializeGroup

      public void initializeGroup()
      Initialize the algorithm for the next feature. (Clears the data containers.)

    • addVariant

      public void addVariant(String variantString)
      Add a variant.
      Parameters:
      variantString - a tab delimited string with variantId,pn,callCopies,nullScore,altScore

    • calculateValues

      public String calculateValues()
      Calculate the p-value and other values for the current genetic feature.
      Returns:
      the values as a tab delimited string