Multiple plot functions

Usage

plot_tsne(
  dat = NULL,
  mod = NULL,
  group = NULL,
  label_group = TRUE,
  position = "right",
  perplexity = 30,
  pca = FALSE,
  ncomp = 70,
  main = "t-SNE",
  size = 1.4,
  shape = 21,
  source = "auto",
  omics = NULL,
  cluster = NULL,
  seed = NULL,
  base_size = 8,
  base_family = "sans",
  ...
)

plot_network(
  dat,
  group = NULL,
  label = FALSE,
  cutoff = NULL,
  layout = igraph::layout_with_fr,
  vertex.size = 5,
  label.dist = NULL,
  edge.width = NULL,
  vertex.frame.color = "white",
  vertex.label.color = "black",
  vertex.label.cex = 0.5,
  vertex.label.dist = 0,
  edge.curved = 0.5,
  vertex.label.degree = -pi/2,
  position = "bottomright",
  source = "auto",
  omics = NULL,
  cluster = NULL,
  seed = NULL,
  ...
)

plot_weights(
  weights,
  plot.which = "all",
  top = 20,
  labels = NULL,
  weight_source = "auto",
  cluster = NULL,
  ncol = NULL,
  base_size = 8,
  base_family = "sans"
)

plot_circos(
  mat,
  names.list = NULL,
  group = NULL,
  cut.off = NULL,
  highlight = NULL,
  source = c("adj_var", "adj_dat"),
  ...
)

plot_embed(dat, group = NULL, position = "bottom", pch = 20, ...)

plot_umap(
  dat,
  group = NULL,
  main = "UMAP",
  label_group = TRUE,
  pca = TRUE,
  ncomp = 70,
  position = "right",
  pch = 21,
  size = 1.4,
  config = umap::umap.defaults,
  method = "naive",
  source = "auto",
  omics = NULL,
  cluster = NULL,
  seed = NULL,
  base_size = 8,
  base_family = "sans",
  ...
)

Arguments

dat: A data frame or matrix
mod: A fitted model object used to compute embeddings internally. Can also be an mrf3_fit object.
group: Class group
label_group: Logical; whether to draw group labels on embeddings.
position: Legend position.
perplexity: tSNE perplexity used by Rtsne::Rtsne().
pca: Logical; whether to run PCA before embedding.
ncomp: Number of principal components retained when pca = TRUE.
main: a main title for the plot, see also title.
size: Point size in scatter plots.
shape: Point shape in embedding plots.
source: Source used when dat/mod/weights is an mrf3_fit. Supported values for mrf3_fit inputs: "auto", "specific_shared", "specific_specific", "clustering", "robust_clustering", "reconstruction_weight_all", "reconstruction_weight_block", "reconstruction_fused_block". plot_network() and plot_circos() additionally accept a pairwise_imd_analysis object with source = "adj_var" or "adj_dat". The t-SNE model path requires a square sample-similarity source; use the dat argument for a raw rectangular feature matrix. For plot_weights(), supported values are: "auto", "imd", "cluster_imd", "mrf".
omics: Optional omics block name when source is block-specific.
cluster: Optional cluster label when source uses cluster-level IMD output.
seed: Optional non-negative integer seed for stochastic embedding or network-layout initialization. The caller's RNG state is restored.
base_size: Base font size used by the multiRF plotting theme.
base_family: Font family used by the multiRF plotting theme.
...: Additional arguments passed to the underlying plotting routine.
label: A logical parameter that determine whether to show labels or not.
cutoff: Edge-weight cutoff. For a regular sample matrix the default is the mean positive upper-triangle weight; for pairwise IMD it is the 75th percentile of positive upper-triangle weights.
layout: Network layout function for graph plotting.
vertex.size: Vertex size in network plots.
label.dist: Label distance in network plots.
edge.width: Edge width in network plots.
vertex.frame.color: Vertex frame color in network plots.
vertex.label.color: Vertex label color in network plots.
vertex.label.cex: Vertex label font size in network plots.
vertex.label.dist: Vertex label distance in network plots.
edge.curved: Edge curvature in network plots.
vertex.label.degree: Vertex label angle in network plots.
weights: Weight vector or weight-carrying object for plot_weights().
plot.which: Which omics block(s) to plot in plot_weights().
top: The number of top weighted variables to show in the plot. The default is 20. Can be chosen from numerical values or all for setting the parameter = NULL.
labels: Optional panel labels in plot_weights().
weight_source: Alias of source used by plot_weights().
ncol: Number of facet columns in plot_weights(). By default, plots with more than two blocks use two columns to preserve readable feature labels.
mat: Adjacency-like matrix for plot_circos().
names.list: Named list of variable groups for plot_circos().
cut.off: Cutoff applied before plotting a circos graph.
highlight: Vector of sectors to highlight in plot_circos().
pch: Point symbol for base plotting functions.
config: UMAP configuration object passed to umap::umap().
method: UMAP backend method passed to umap::umap().

Value

plot_tsne(), plot_umap(), and plot_weights() return a ggplot object. plot_network() invisibly returns an igraph object, plot_circos() invisibly returns the chord-diagram result, and the deprecated plot_embed() invisibly returns NULL after drawing.