fitNullModelGSA

Fit a null model for a mixed linear model gene set analysis. The correlation matrix used to model gene-gene correlations can be generated using the buildCorMatrix method.

fitNullModelGSA(
  object,
  cormatrix = NULL,
  covar = NULL,
  Zcutoffs = NULL,
  INT = FALSE,
  method = c("GENESIS"),
  ...
)

# S4 method for rvatResult
fitNullModelGSA(
  object,
  cormatrix = NULL,
  covar = NULL,
  Zcutoffs = NULL,
  INT = FALSE,
  method = c("GENESIS"),
  ...
)

Arguments

object

rvbResult-class object

cormatrix

a correlation matrix with row and column names corresponding to the units in the rvbResult. Can be a sparse matrix. See buildCorMatrix.

covar

a vector of covariates to include when fitting the null model.

Zcutoffs

A vector (length=2, minimum and maximum) of cutoffs to apply to the Z-scores. Z scores below/above these cutoffs will be set equal to the cutoff.

INT

Apply inverse normal transformation to Z-scores? Defaults to FALSE.

method

Method to use to fit the mixed linear model, currently 'GENESIS' is implemented.

...

Additional arguments to pass to GENESIS::fitNullModel