This function creates a list of class ConsensusOptions that holds options for consensus calculations.
This list holds options for a single-level analysis.
newConsensusOptions(
calibration = c("full quantile", "single quantile", "none"), # Simple quantile scaling options
calibrationQuantile = 0.95,
sampleForCalibration = TRUE,
sampleForCalibrationFactor = 1000,
# Consensus definition
consensusQuantile = 0,
useMean = FALSE,
setWeights = NULL,
suppressNegativeResults = FALSE,
# Name to prevent files clashes
analysisName = "")
Calibration method. One of "full quantile", "single quantile", "none"
(or a unique abbreviation of one of them).
if calibration is "single quantile",
input data to a consensus calculation
will be scaled such that their calibrationQuantile quantiles will agree.
if TRUE, calibration quantiles will be determined from a sample of
network
similarities. Note that using all data can double the memory footprint of the function and the function
may fail.
Determines the number of samples for calibration: the number is
1/calibrationQuantile * sampleForCalibrationFactor. Should be set well above 1 to ensure accuracy of
the sampled quantile.
Quantile at which consensus is to be defined. See details.
Logical: should the consensus be calculated using (weighted) mean rather than a quantile?
Optional specification of weights when useMean is TRUE.
Logical: should negative consensus results be replaced by 0? In a typical network
connstruction, negative topological overlap values may results with TOMType = "signed Nowick".
Optional character string naming the consensus analysis. Useful for identifying partial consensus calculation in hierarchical consensus analysis.
A list of type ConsensusOptions that holds copies of the input arguments.