MetaDE.pvalue: Meta analysis by combining p-value

In metaOmic/MetaDE: MetaDE: Transcriptome meta-analysis for differentially expressed gene detection

Description

Meta analysis by combining p-value The MetaDE is a function to identify genes associated with the response/phenoype of interest (can be either group, continuous or survival) by combining p-values from multiple studies(datasets). The main input consists of p-values from your own method/calculations.

Usage

MetaDE.pvalue(x, meta.method, rth = NULL, parametric = TRUE)

Arguments

x	is a list with components: p: a list of p values for each dataset. bp: a list of p values calculated from permutation for each dataset. This part can be NULL if you just have the p-values from your own method.
meta.method	is a character to specify the Meta-analysis method used to combine the p-values.
rth	is the option for roP and roP.OC method. rth means the rth smallest p-value.
parametric	is a logical values indicating whether the parametric methods is chosen to calculate the p-values in meta-analysis.
x	is a list with components:

Value

a list with components:

• stat: a matrix with rows representing genes. It is the statistic for the selected meta analysis method of combining p-values.

• pval: the p-value from meta analysis for each gene for the above stat.

• FDR: the FDR of the p-value for each gene for the above stat.

• AW.weight: The optimal weight assigned to each dataset/study for each gene if the 'AW' method was chosen.

Examples

data('Leukemia')
data('LeukemiaLabel')
data <- Leukemia
K <- length(data)
clin.data <- lapply(label, function(x) {data.frame(x)} )
for (k in 1:length(clin.data)){
 colnames(clin.data[[k]]) <- "label"
}
select.group <- c('inv(16)','t(15;17)')
ref.level <- "inv(16)"
data.type <- "continuous"
ind.method <- c('limma','limma','limma')
resp.type <- "twoclass"
paired <- rep(FALSE,length(data))
ind.res <- Indi.DE.Analysis(data=data,clin.data= clin.data, 
                        data.type=data.type,resp.type = resp.type,
                        response='label',
                        ind.method=ind.method,select.group = select.group,
                        ref.level=ref.level,paired=paired)
meta.method <- "AW"
meta.res <- MetaDE.pvalue(ind.res,meta.method,rth=NULL,parametric=TRUE)
summary <- data.frame(ind.p = meta.res$ind.p,
                      stat = meta.res$meta.analysis$stat,
                      pval = meta.res$meta.analysis$pval,
                      FDR = meta.res$meta.analysis$FDR,
                      weight = meta.res$meta.analysis$AW.weight)

metaOmic/MetaDE documentation built on May 22, 2019, 6:54 p.m.