plotSim {adjclust}R Documentation

Plot (dis)similarity matrix

Description

Heatmap of the (dis)similarity matrix

Usage

plotSim(
  mat,
  type = c("similarity", "dissimilarity"),
  clustering = NULL,
  dendro = NULL,
  palette = heat.colors,
  breaks = 10,
  log = TRUE,
  h = NULL,
  stats = c("R.squared", "D.prime"),
  main = NULL,
  col.clust = "darkblue",
  lwd.clust = 2,
  xaxis = FALSE,
  naxis = 10
)

Arguments

mat

matrix to plot. It can be of class 'matrix', 'dgCMatrix', 'dsCMatrix', 'dist', 'HTCexp', 'snpMatrix'.

type

input matrix type. Can be either "similarity" or "dissimilarity" (kernels are supposed to be of type "similarity").

clustering

vector of length the number of rows (columns) of the matrix that contains a contiguity constrained clustering (as provided by select for instance). If supplied the clustering is superimposed over the heatmap.

dendro

chac object as provided, e.g., by the function adjClust (or any of the other wrappers).

palette

color palette. Default to heat.colors

breaks

number of breaks used to set colors from the palette. Those are based on the quantiles of the matrix entries and for skewed distributions the actual number used to set the palette can be lower than breaks.

log

logical. Should the breaks be based on log-scaled values of the matrix entries. Default to TRUE.

h

if mat is of class "snpMatrix", band parameter used to compute the linkage desiquilibrium (see ld).

stats

if mat is of class "snpMatrix", type of linkage desiquilibrium measure (see ld).

main

graphic title.

col.clust

color for the borders of the clusters (if clustering is provided).

lwd.clust

line width for the borders of the clusters (if clustering is provided).

xaxis

logical. Should a x-axis be displayed? Default to FALSE

naxis

number of breaks to display on the x-axis. For HTCexp objects, the axis is displayed in terms of Mpb and for the other types of input, it is displayed in terms of bin number. Default to 10.

Details

This function produces a heatmap for the used (dis)similarity matrix that can be used as a diagnostic plot to check the consistency between the obtained clustering and the original (dis)similarity

See Also

select, adjClust

Examples

# input as HiTC::HTCexp object
## Not run: 
if (require("HiTC", quietly = TRUE)) {
  load(system.file("extdata", "hic_imr90_40_XX.rda", package = "adjclust"))
  plotSim(hic_imr90_40_XX)
  
  # with a constrained clustering
  res <- hicClust(hic_imr90_40_XX, log = TRUE)
  selected.capushe <- select(res)
  plotSim(hic_imr90_40_XX, clustering = selected.capushe, xaxis = TRUE)
  plotSim(hic_imr90_40_XX, clustering = selected.capushe, dendro = res)
}
## End(Not run)

plotSim(dist(iris[ ,1:4]), log = FALSE)

[Package adjclust version 0.5.99 Index]