CombP: Convergent evidence based on combined p-values.

In KanduriC/GenRank: Candidate gene prioritization based on convergent evidence

Description

CombP returns ranks of the genes based on p-values combined using the famous 'fisher' or 'z-transform' methods.

Usage

CombP(file, weight, method = c("fisher", "z.transform", "logit"),
  na.remove = FALSE)

Arguments

file	A tab-delimited text file with a minimum of 3 columns. First column should contain gene names, second column should indicate the evidence type and third column should contain non-negative numeric values (p-values).
weight	A numeric vector containing weights of evidence types. For example, sample sizes of various evidence types. If not provided, equal weight is given to all evidence types.
method	A character string among 'fisher', 'z.transform' or 'logit'.
na.remove	An optional argument, defaults to FALSE. Set this argument to TRUE if all the genes were not detected across all evidence types.

Value

If all the inputs are in the correct format as suggested, then the output will be a dataframe containg genes, their combined p-values and corresponding ranks.

Examples

cus.weights <- c(100,50,200,300,150,400)
input_file_P <- system.file("extdata","CombP_toydata.txt",package="GenRank")
CP_ranking <- CombP(input_file_P, method = "fisher", na.remove = TRUE)
CP_ranking_z <- CombP(input_file_P, method = "z.transform", na.remove = TRUE, weight = cus.weights)

KanduriC/GenRank documentation built on May 8, 2019, 4:42 p.m.