msc.pca: Prinicple Component Analysis based on MSC

In GeertsManon/rKOMICS: Minicircle Sequence Cluster (MSC) Analyses

Description

The msc.pca allows you to perform Principle Component Analysis to summarize MSCs variation in all samples or in a subset of samples.

Usage

msc.pca(clustmatrix, samples, groups, n = 20, labels = TRUE, title = NULL)

Arguments

clustmatrix	a cluster matrix
samples	a vector containing the names of the samples. This can include all samples or it can be a subset.
groups	a vector specifying to which group (e.g. species) the samples belong to.
n	number of clusters to select with highest contribution to PCA.
labels	a logical parameter indicating whether to use labels on the PCA plot or not. Default is set to true.
title	the title of the graph

Value

plot	a PCA plot.
eigenvalues	a barplot showing percentage of explained variances.
clustnames	list of cluster names with highest contribution to PCA.

Examples

data(matrices)
data(exData)

### run function with all samples
res.pca <- lapply(matrices, function(x) msc.pca(x, samples = exData$samples, 
                  groups = exData$species, n=30, labels=FALSE, title=NULL))

res.pca$id93$eigenvalues
res.pca$id93$plot

### use clusters with highest contribution to visualize in a heatmap
msc.heatmap(matrices[["id93"]][res.pca$id93$clustnames,], samples = exData$samples,
            groups = exData$species)

### run function with a subset of samples
### you will be asked to confirm
table(exData$species)
hybrid <- which(exData$species=="hybrid")
# pca.subset <- msc.pca(clustmatrix = matrices[["id97"]], 
#                       samples = exData$samples[hybrid], 
#                       groups = exData$species[hybrid], labels = TRUE, 
#                       title = "PCA only with hybrids")

GeertsManon/rKOMICS documentation built on Dec. 17, 2021, 9:28 p.m.