summarize.top.clusters: Summarize the top clusters

In tamartsi/Aclust: Clustering functions and analysis pipeline for methylation outcome data based on clustering and GEEs

Description

Returns information about the top cluster in the GEE analysis

Usage

summarize.top.clusters(betas, covariates, exposure, id, clusters.GEE.results = NULL, clusters.GEE.results.file = NULL, minimum.sites = 2, top.number = 10, cutoff.fdr.pval = 0.05, cutoff.effect.size = NULL, annot = NULL, annotation.file.name = NULL, required.annotation = c("IlmnID", "Coordinate_37", "UCSC_RefGene_Name", "UCSC_RefGene_Group", "UCSC_CpG_Islands_Name", "Relation_to_UCSC_CpG_Island"), file.to.print.report = NULL)

Arguments

betas	An (m by n) matrix of methylation values of $m$ methylation sites measured on $n$ individuals
covariates	Either a (n by p) vectors of adjusting covariates, ordered so that its rows corresponds to the columns of the ‘betas’ matrix, or NULL if there are no covariates.
exposure	A vector of size $n$ of exposure values for each individual. Same order as the covariates.
id	A vector of size $n$ of the IDs of the individual. The IDs should match the column names of the matrix ‘betas’.
clusters.GEE.results	A matrix of results from a GEE analysis of clusters.
clusters.GEE.results.file	A file with the results from a GEE analysis of clusters (if clusters.GEE.results is not given)
minimum.sites	A minimum number of sites in a cluster to be considered in the analysis.
top.number	The required number of top clusters to be reported.
cutoff.fdr.pval	A significance p-value threshold (after FDR correction). Only clusters with p-value smaller than cutoff.fdr.pval will be reported.
cutoff.effect.size	An effect size threshold. Only clusters with estimated effect size larger than cutoff.effect.size will be reported.
annot	annotation data table. The package uses the Illumina annotation file “illumina_450_manifest_v.1.2".
annotation.file.name	A name of annotation file to read. By default it is not required, since one can use ‘annot’, which is in the package.
required.annotation	What is the annotation to be reported on each of the sites in the reported clusters?
file.to.print.report	File name to print tables of annotation, effect sizes, and individual site analyses of sites in the top clusters.
print.progress	Print status messages while the function progresses.

Value

top.clusters	A data frame summarizing the GEE results (exposure effect estimates, etc) of the top clusters
annotation.top.clusters	Annotations of the sites from the top clusters
individual.sites.analysis	Results of individual sites analysis, of the sites in the top clusters.

Note

Individual sites analysis uses the robust standard error.

Examples

data(betas.7)
exposure <- rbinom(ncol(betas.7), 1,prob = 0.5) ## generate random exposure
covariates <- matrix(rnorm(2*ncol(betas.7)), ncol = 2)
rownames(covariates) <- colnames(betas.7)

## not run! to create new annotation file from the 
## IlluminaHumanMethylation450k.db package:
## annot <- create.annot.triche(rownames(betas.7)

data(annot.7)

clusters.list <- assign.to.clusters(betas.7, annot.7)
GEE.results.clusters <- GEE.clusters(betas.7, clusters.list, exposure, covariates, id = colnames(betas.7), working.cor = "ex")
top.clusters.summary <- summarize.top.clusters(betas.7, covariates, exposure, id = colnames(betas.7), GEE.results.clusters, "results.tex", annot= annot.7)

tamartsi/Aclust documentation built on May 31, 2019, 2:56 a.m.