Class DifferentialExpressionAnalysisUtil

java.lang.Object
ubic.gemma.core.analysis.expression.diff.DifferentialExpressionAnalysisUtil
public class DifferentialExpressionAnalysisUtil extends Object
A helper class for the differential expression analyzers. This class contains helper methods commonly needed when performing an analysis.
Author:
keshav

  • Constructor Details

    • DifferentialExpressionAnalysisUtil

      public DifferentialExpressionAnalysisUtil()

  • Method Details

    • blockComplete

      public static boolean blockComplete(BioAssaySet expressionExperiment, Collection<ExperimentalFactor> factors)
      Returns true if the block design is complete and there are at least 2 biological replicates for each "group", false otherwise. When determining completeness, a biomaterial's factor values are only considered if they are equivalent to one of the input experimental factors.
      Parameters:
      expressionExperiment - the experiment
      factors - to consider completeness for.
      Returns:
      true if block complete

    • blockComplete

      public static boolean blockComplete(Collection<BioMaterial> samples, Collection<ExperimentalFactor> factors)
      A variant that judges from a given set of samples rather than every sample in the experiment. Pass the samples that will actually be modelled, so that DE_Exclude and outlier samples do not make a design look block-complete when the fitted set is not: the interaction term chosen from the full set is unestimable on the smaller one.

      The sibling checkValidForLm(Collection, ExperimentalFactor) exists for the same reason.

    • checkValidForLm

      public static boolean checkValidForLm(BioAssaySet expressionExperiment, ExperimentalFactor experimentalFactor)
      Check that the factorValues are measurements, or that there are at least two assays for at least one factor value. Otherwise the model fit will be perfect and pvalues will not be returned.
      Parameters:
      expressionExperiment - the experiment
      experimentalFactor - exp. factor
      Returns:
      true if it's okay, false otherwise.

    • checkValidForLm

      public static boolean checkValidForLm(Collection<BioMaterial> samples, ExperimentalFactor experimentalFactor)
      A variant that judges from a given set of samples rather than every sample in the experiment. Pass the samples that will actually be modelled, so that DE_Exclude and outlier samples do not make a factor look analyzable.