我知道R中的p.adjust函数,它能很好地满足我的需要。但是,现在我想根据FDR (Benjamini & Hochberg)方法来修正significance threshold (alpha)而不是p-values本身。例如,我们有一个10的原始p值:
0.0001,0.001,0.024,0.56,0.0077,0.55,0.0025,0.01,0.015,1在Bonferroni的情况下,这非常简单:
alpha_Bonferroni_corrected = 0.01/ number of tests (10 in our example)=0.001但对于FDR来说,这将是一个更棘手的问题。在R中有这样的函数吗?
发布于 2018-11-11 17:06:59
mutoss包似乎提供了更大的灵活性
library(mutoss)
alpha <- 0.01
set.seed(1234)
p <-c(runif(10, min=0, max=0.01), runif(10, min=0.9, max=1))
result <- adaptiveBH(p, alpha)
resulthttps://stackoverflow.com/questions/53243979
复制相似问题