dipterix/focr: A False Overlapped-Cluster Rate Control ('FOCR') Framework

Provides a two-stage framework for multiple testing problems with topological constraints. In the first stage, clusters of hypotheses are selected as regions of interests. The clusters can be overlapped. The false overlapped-cluster rate w.r.t. the clusters is controlled. In the second stage, conditional p-values are calculated at the individual level in a post-selection inference fashion. The FDR controlling procedures are applied to these conditional p-values. Two functions are proposed: 'focr_initial' and 'focr'. The function 'focr' provides a bundled stage-I and stage-II procedure with preset settings. The function 'focr_initial' only controls the stage-I 'FOCR' and calculate the conditional p-values. However, it allows more flexible controls on the clusters.

Getting started

Package details

Maintainer
LicenseMIT + file LICENSE
Version0.1.0
URL http://dipterix.github.io/focr
Package repositoryView on GitHub
Installation Install the latest version of this package by entering the following in R:
install.packages("remotes")
remotes::install_github("dipterix/focr")
dipterix/focr documentation built on Dec. 20, 2021, 12:03 a.m.