hcsig: hcsig: hcluster Significance

In jlaffy/statistrics: Basic Statistics For Single-Cell (or bulk) RNA Analyses

Description

hcluster Significance. For clusters derived from hierarchical clustering (with hclust), data is retrieved.

Usage

hcsig(k, mat, fc.value = 3, p.value = 10^(-4), p.value.2 = 10^(-5),
  pval.adjust = NULL, reorder = TRUE, fc.sort = T, pval.sort = F,
  returning = "all")

Arguments

k	a list of character vectors; sets of cell members belonging to each cluster.
mat	a matrix of gene expression data (cells by genes)
fc.value	fold change value below which differential gene expression is deemed insignificant.
p.value	p-value above which differential gene expression is deemed insignificant.
p.value.2	p-value above which differential gene expression is deemed insignificant with higher cutoff (sig.2)
pval.adjust	NULL or character string. If NULL, do not adjust p-values. If string, adjust p-values using the method specified.
reorder	if TRUE, the list of clusters is reordered by most to least significant.
fc.sort	if TRUE, significantly differentially expressed genes are sorted by fold change (highest first). Default is TRUE.
pval.sort	if TRUE, significantly differentially expressed genes are sorted by p.value (highest first). pval.sort=TRUE overrides fc.sort=TRUE. Default is FALSE.
returning	return one of p-values, fold changes, or both from call of DEgenes() to sig().

Value

list of length 4. Each object in the list is also a list. Each list has the same length, which is the length of k arg (the number of clusters). The lists are list$k, same as input; list$sig.1, the significant genes' p-values for each cluster; list$sig.2, list$sig.1 filtered such that only genes with p value higher than p.value.2 are included. list$sig.3, list$sig.1 filtered such that each gene only appears once across the clusters, wherever it had the highest p-value.

jlaffy/statistrics documentation built on May 23, 2019, 4:04 a.m.