extractClusters: Extract the clusters from a Clustered Heatmap

In axm323/dataVisEasy: datavisEasy: Data visualizations Using Annotations Made Easy

Extract the clusters from a Clustered Heatmap

Description

Provided a clustered heatmap, and a number of groups the dendrogram should be split into, will return a vector detailing which genes/samples are in each cluster. Particularly helpful after setting row.groups or column.groups in the heatmap functions and the groupings of the genes or samples in these groups need to be identified.

Usage

extractClusters(data, heatmap,to.extract = "genes",nclusters, GeneGroup_Name = NULL,
  SampleGroup_Name = NULL, all.genes = NULL, all.samples = NULL)

Arguments

data	numeric data matrix with samples/observations in the columns and genes/variables in the rows. Should be the same data supplied to the heatmap function
heatmap	Output of a heatmap generated from myHeatmap or myHeatmapByAnnotation. The heatmap generated from these functions can be saved to a variable which should serve as the input here. IMPORTANT. The heatmap supplied here MUST have show.colnames and show.rownames = TRUE.
to.extract	whether clusters should be extracted for genes (rows), samples(columns), or both. Can accept "genes", "samples", or "both"
nclusters	numeric, how many clusters should be extracted. If to.extract is set to "both", numeric vector should be supplied listing number of gene (row clusters) first
GeneGroup_Name	optional name for the groupings to be extracted
SampleGroup_Name	optional name for the groupings to be extracted
all.genes	gene names, if annotations should be in placed in larger gene list than matrix in question. Default assumes that the rownames of the input data represent all genes in question (this is true even if the genes in the heatmap are a subset of the input data). Genes not represented in the heatmap or input data that do not have a cluster designation will be assigned "No_Annot".
all.samples	sample names, if annotations should be in placed in larger sample list than matrix in question. Default assumes that the colnames of the input data represent all samples in question. If these annotations should be placed within a larger list (for example if the samples in the input data are a subset of all samples). Samples not represented in the heatmap or input data that do not have a cluster designation will be assigned "No_Annot".

Details

This function is helpful for assigned annotations to samples or genes based on how they cluster. The options of all.samples and all.genes allow these cluster assignments to be placed within a larger list of variables (useful if combining these annotations with other pre-existing annotations, where the vectors need to be of the same length to merge correctly).

If GeneGroup_Name and/or SampleGroup_Name are provided, these vectors can be passed directly to update_annotations or update_annotations.genes

Value

a named vector or a list object of two named vectors (in a case where both genes and samples are extracted) where the assignments of each gene/sample are designated by letter A:num.groups. Importantly, these groupings are assigned as they appear in the heatmap itself where cluster assignment A is located on the top left (going across from left to right in the case of extractSamples and going down from top to bottom in the case of extractGenes)

Author(s)

~~Alison Moss~~

See Also

See set_annotations and update_annotations.genes or update_annotations and update_annotations.genes for how the output of this function can be passed directly to these functions to set or update annotations stored in the params list object

Examples

##initiate parameters
initiate_params()

##create clustered heatmap, defining how many groups clusters should split into
heatmap <- myHeatmap(RAGP_norm, row.groups = 6, col.groups = 5)

##extract clusters for rows, named from top to bottom
row.clusts <- extractClusters(RAGP_norm, heatmap = heatmap, to.extract = "genes", nclusters = 6)
head(row.clusts)

##extract clusters for columns, named from left to right
col.clusts <- extractClusters(RAGP_norm, heatmap = heatmap, to.extract = "samples", nclusters = 5)
head(col.clusts)

##extract both
all.clusts <- extractClusters(RAGP_norm, heatmap = heatmap, to.extract = "both", nclusters = c(6,5))
head(all.clusts$Gene_Groups)
head(all.clusts$Sample_Groups)



###extract clusters from a heatmap with subsetted genes, catecholaminergic/cholinergic receptors
#subset data for receptors
RAGP_norm.receptors <- subsetGenes(RAGP_norm, list = c("Chrm","Chrn","Adr"),
exact = FALSE, subset.by = "name")
heatmap.receptors <- myHeatmap(RAGP_norm.receptors,
  main = "Catecholaminergic and Cholinergic Receptors", row.groups = 6)

receptor.groups <- extractClusters(RAGP_norm.receptors, heatmap = heatmap.receptors,
  to.extract = "genes", nclusters = 6, GeneGroup_Name = "Receptor_Groups")
head(receptor.groups)

##if we want the annotations to include genes not present in the heatmap, set all.genes
###genes not present in the heatmap will be assigned "No_Annot" for no annotation
receptor.groups1 <- extractClusters(RAGP_norm.receptors, heatmap = heatmap.receptors,
  to.extract = "genes", nclusters = 6, GeneGroup_Name = "Receptor_Groups",
  all.genes = rownames(RAGP_norm))
head(receptor.groups1)

##the same result would be true if a list was supplied to the heatmap to subset
heatmap.receptors2 <- myHeatmap(RAGP_norm, list = c("Chrm","Chrn","Adr"), exact = FALSE,
  main = "Catecholaminergic and Cholinergic Receptors", row.groups = 6)
receptor.groups2 <- extractClusters(RAGP_norm, heatmap = heatmap.receptors2,
  to.extract = "genes", nclusters = 6, GeneGroup_Name = "Receptor_Groups")
head(receptor.groups2)

axm323/dataVisEasy documentation built on Feb. 1, 2024, 11:53 p.m.