runPipeline: Perform differential methylation analysis

In MEAL: Perform methylation analysis

Description

Wrapper for analysing differential methylation and expression at region and probe level.

Usage

runPipeline(
  set,
  variable_names,
  covariable_names = NULL,
  model = NULL,
  weights = NULL,
  num_vars,
  sva = FALSE,
  betas = TRUE,
  range,
  analyses = c("DiffMean"),
  verbose = FALSE,
  warnings = TRUE,
  DiffMean_params = NULL,
  DiffVar_params = list(coefficient = 1:2),
  rda_params = NULL,
  method = "ls"
)

Arguments

set	GenomicRatioSet, eSet derived object or SummarizedExperiment
variable_names	Character vector with the names of the variables that will be returned as result.
covariable_names	Character vector with the names of the variables that will be used to adjust the model.
model	Model matrix or formula to get model matrix from set.
weights	weights used in the lmFit model (default NULL)
num_vars	Numeric with the number of variables in the matrix for which the analysis will be performed. Compulsory if equation is not null.
sva	Logical. Should Surrogate Variable Analysis be applied? (Default: FALSE)
betas	If set is a GenomicRatioSet, should beta values be used? (Default: TRUE)
range	GenomicRanges with the region used for RDA
analyses	Vector with the names of the analysis to be run (DiffMean and/or DiffVar).
verbose	Logical value. If TRUE, it writes out some messages indicating progress. If FALSE nothing should be printed.
warnings	Should warnings be displayed? (Default:TRUE)
DiffMean_params	List with other parameter passed to runBumphunter function.
DiffVar_params	List with other parameter passed to runBumphunter function.
rda_params	List with other parameter passed to runRDA function.
method	String indicating the method used in the regression: "ls" or "robust". (Default: "ls")

Details

This function is the main wrapper of the package. First, it simplifies the the set to only contain the common samples between phenotype and features. In addition, it allows to change the class of the variables and to apply genomic models (more information on preparePhenotype). Afterwards, analysis per probe and per region are done merging the results in an AnalysisResults object.

Default linear model will contain a sum of the variables and covariables. If interactions are desired, a costum formula can be specified. In that case, variables and covariables must also be specified in order to assure the proper work of the resulting AnalysisResult. In addition, the number of variables of the model for which the calculation will be done must be specified.

Value

ResultSet object

Examples

if (require(minfiData)){
set <- ratioConvert(mapToGenome(MsetEx[1:10,]))
 res <- runPipeline(set, variable_names = "Sample_Group") 
 res
}

MEAL documentation built on Feb. 3, 2021, 2 a.m.