Package 'colorrepel'

Average expression values per cluster

Description

Average expression values per cluster

Usage

average_clusters(
  mat,
  metadata,
  cluster_col = "cluster",
  if_log = TRUE,
  cell_col = NULL,
  low_threshold = 0,
  method = "mean",
  output_log = TRUE,
  cut_n = NULL
)

Arguments

mat	expression matrix
metadata	data.frame or vector containing cluster assignments per cell. Order must match column order in supplied matrix. If a data.frame provide the cluster_col parameters.
cluster_col	column in metadata with cluster number
if_log	input data is natural log, averaging will be done on unlogged data
cell_col	if provided, will reorder matrix first
low_threshold	option to remove clusters with too few cells
method	whether to take mean (default), median, 10% truncated mean, or trimean, max, min, sum
output_log	whether to report log results
cut_n	set on a limit of genes as expressed, lower ranked genes are set to 0, considered unexpressed

Value

average or other desired calculation by group/cluster matrix

Examples

mat <- average_clusters(data.frame(
  z = c(1, 2, 3, 4, 5, 6),
  y = c(1, 2, 3, 4, 5, 6),
  x = c(1, 2, 3, 4, 5, 6)
), metadata = c(1, 1, 2), method = "sum")

---

Rowwise math from matrix/data.frame per cluster based on another vector/metadata, similar to clustifyr::average_clusters but ids as rows

Description

Rowwise math from matrix/data.frame per cluster based on another vector/metadata, similar to clustifyr::average_clusters but ids as rows

Usage

average_clusters_rowwise(
  mat,
  metadata,
  cluster_col = "cluster",
  if_log = FALSE,
  cell_col = NULL,
  low_threshold = 0,
  method = "mean",
  output_log = FALSE,
  cut_n = NULL,
  trim = FALSE
)

Arguments

mat	expression matrix
metadata	data.frame or vector containing cluster assignments per cell. Order must match column order in supplied matrix. If a data.frame provide the cluster_col parameters.
cluster_col	column in metadata with cluster number
if_log	input data is natural log, averaging will be done on unlogged data
cell_col	if provided, will reorder matrix first
low_threshold	option to remove clusters with too few cells
method	whether to take mean (default), median, 10% truncated mean, or trimean, max, min, sum
output_log	whether to report log results
cut_n	set on a limit of genes as expressed, lower ranked genes are set to 0, considered unexpressed
trim	whether to remove 1 percentile when doing min caluculation

Value

average expression matrix, with genes for row names, and clusters for column names

Examples

mat <- average_clusters_rowwise(data.frame(
  y = c(1, 2, 3, 4, 5, 6),
  x = c(1, 2, 3, 4, 5, 6)
), metadata = c(1, 2, 1, 2, 1, 2), method = "min")

---

Balanced downsampling of matrix/data.frame based on cluster assignment vector

Description

Balanced downsampling of matrix/data.frame based on cluster assignment vector

Usage

by_cluster_sampling(df, vec, frac, seed = 34)

Arguments

df	expression matrix or data.frame
vec	vector of ids
frac	fraction 0-1 to downsample to
seed	sampling randomization seed

Value

list with new downsampled matrix/data.frame and id vector

Examples

res <- by_cluster_sampling(data.frame(y = c(1, 2, 3, 4, 5, 6)),
  vec = c(1, 2, 1, 2, 1, 2), frac = 0.5
)

---

Distance calculations for spatial coord

Description

Distance calculations for spatial coord

Usage

calc_distance(
  coord,
  metadata,
  cluster_col = "cluster",
  collapse_to_cluster = FALSE
)

Arguments

coord	dataframe or matrix of spatial coordinates, cell barcode as rownames
metadata	data.frame or vector containing cluster assignments per cell. Order must match column order in supplied matrix. If a data.frame provide the cluster_col parameters.
cluster_col	column in metadata with cluster number
collapse_to_cluster	instead of reporting min distance to cluster per cell, summarize to cluster level

Value

min distance matrix

---

Reorder ggplot colors to maximize color differences in space

Description

Reorder ggplot colors to maximize color differences in space

Usage

color_repel(
  g,
  coord = NULL,
  groups = NULL,
  nsamp = 50000,
  sim = NULL,
  severity = 0.5,
  verbose = FALSE,
  downsample = 5000,
  polychrome_recolor = FALSE,
  seed = 34,
  col = "colour",
  autoswitch = TRUE,
  layer = 1,
  out_orig = FALSE,
  out_worst = FALSE,
  ggbuild = NULL
)

Arguments

g	ggplot plot object
coord	coordinates, default is inferred
groups	groups corresponding to color/fill, default is inferred
nsamp	how many random sampling color combinations to test, default 50000
sim	passing a colorbind simulation function if needed
severity	severity of the color vision defect, between 0 and 1
verbose	whether to print messages
downsample	downsample when too many datapoints are present, or use chull
polychrome_recolor	whether to replace the original colors with polychrome creation
seed	sampling randomization seed
col	colour or fill in ggplot
autoswitch	try to switch between colour and fill automatically
layer	layer to detect color, defaults to first
out_orig	output the original colors as named vector
out_worst	output the worst combination instead of best
ggbuild	already built ggplot_built object if available

Value

vector of reordered colors

Examples

a <- ggplot2::ggplot(ggplot2::mpg, ggplot2::aes(displ, hwy)) +
  ggplot2::geom_point(ggplot2::aes(color = as.factor(cyl)))
new_colors <- color_repel(a)
b <- a + ggplot2::scale_color_manual(values = new_colors)

---

Extract custom labels from ggplot object

Description

Extract custom labels from ggplot object

Usage

get_labs(g, ggbuild = NULL)

Arguments

g	ggplot object
ggbuild	already built ggplot_built object if available

Value

named vector of labels

Examples

a <- ggplot2::ggplot(ggplot2::mpg, ggplot2::aes(displ, hwy)) +
  ggplot2::geom_point(ggplot2::aes(color = as.factor(cyl))) +
  ggplot2::geom_text(ggplot2::aes(label = model))
get_labs(a)

---

Wrapper to reorder ggplot colors to maximize color differences in space

Description

Wrapper to reorder ggplot colors to maximize color differences in space

Usage

gg_color_repel(
  g = ggplot2::last_plot(),
  col = "colour",
  sim = NULL,
  severity = 0.5,
  verbose = FALSE,
  downsample = 5000,
  nsamp = 50000,
  polychrome_recolor = FALSE,
  seed = 34,
  autoswitch = TRUE,
  layer = 1,
  out_orig = FALSE,
  out_worst = FALSE,
  repel_label = FALSE,
  encircle = FALSE,
  encircle_alpha = 0.25,
  encircle_expand = 0.02,
  encircle_shape = 0.5,
  encircle_threshold = 0.01,
  encircle_nmin = 0.01,
  mascarade = FALSE,
  ggbuild = NULL,
  ...
)

Arguments

g	ggplot plot object
col	colour or fill in ggplot
sim	passing a colorbind simulation function if needed
severity	severity of the color vision defect, between 0 and 1
verbose	whether to print messages
downsample	downsample when too many datapoints are present
nsamp	how many random sampling color combinations to test, default 50000
polychrome_recolor	whether to replace the original colors with polychrome creation
seed	sampling randomization seed
autoswitch	try to switch between colour and fill automatically
layer	layer to detect color, defaults to first
out_orig	output the original colors as named vector
out_worst	output the worst combination instead of best
repel_label	whether to add centroid labels with ggrepel
encircle	whether to draw geom_encircle by cluster
encircle_alpha	alpha argument passed to geom_encircle
encircle_expand	expand argument passed to geom_encircle
encircle_shape	shape/smoothing argument passed to geom_encircle
encircle_threshold	threshold for removing outliers
encircle_nmin	number of near neighbors for removing outliers
mascarade	use mascarade package to outline clusters
ggbuild	already built ggplot_built object if available
...	passed to repel_label

Value

new ggplot object

Examples

a <- ggplot2::ggplot(ggplot2::mpg, ggplot2::aes(displ, hwy)) +
  ggplot2::geom_point(ggplot2::aes(color = as.factor(cyl)))
b <- gg_color_repel(a, col = "colour")

---

Prepare ggplot object to ggplotly-compatible layer and image layer

Description

Prepare ggplot object to ggplotly-compatible layer and image layer

Usage

ggplotly_background(
  g,
  repel_color = TRUE,
  repel_label = TRUE,
  encircle = FALSE,
  mascarade = FALSE,
  width = 5,
  height = 5,
  filename = "temp.png",
  draw_box = NULL,
  background = NULL,
  background_alpha = 1,
  use_cairo = FALSE,
  label_lim = 0.05,
  ggbuild = NULL,
  crop = TRUE,
  size_nudge = 0,
  ...
)

Arguments

g	ggplot plot object
repel_color	whether to rearrange colors
repel_label	whether to add centroid labels with ggrepel
encircle	whether to draw geom_encircle by cluster
mascarade	use mascarade package to outline clusters
width	plot width
height	plot height
filename	temp file location for saving image
draw_box	if a colored background should be included
background	if specified, use this ggplot object or file as background instead
background_alpha	alpha value of background image
use_cairo	whether to use cairo for saving plots, maybe needed for certain ggplot extensions
label_lim	whether to limit labels to avoid edge fraction
ggbuild	already built ggplot_built object if available
crop	whether to call cropping of the background image to remove whitespace
size_nudge	slight image size adjustment, default to none
...	arguments passed to gg_color_repel

Value

plotly object with background image of layers unsupported by plotly

Examples

a <- ggplot2::ggplot(ggplot2::mpg, ggplot2::aes(displ, hwy)) +
  ggplot2::geom_point(ggplot2::aes(color = as.factor(cyl)))
new_colors <- color_repel(a)
b <- ggplotly_background(a, filename = NULL)

---

ggrepel labeling of clusters

Description

ggrepel labeling of clusters

Usage

label_repel(
  g,
  group_col = "auto",
  x = "x",
  y = "y",
  txt_pt = 3,
  remove_current = "auto",
  layer = "auto",
  ggbuild = NULL,
  ...
)

Arguments

g	ggplot object or data.frame
group_col	column name in data.frame, default to "label" or "group" in ggplot data
x	column name in data.frame for x
y	column name in data.frame for y
txt_pt	text size
remove_current	whether to remove current text
layer	text layer to remove, defaults to last
ggbuild	already built ggplot_built object if available
...	arguments passed to geom_text_repel

Value

function, if data.frame input, or new ggplot object

Examples

g <- label_repel(ggplot2::ggplot(mtcars, ggplot2::aes(x = hp, y = wt, color = as.character(cyl))) +
  ggplot2::geom_point(), remove_current = FALSE)

---

Score matrix distances

Description

Score matrix distances

Usage

matrix2_score(dist1, dist2)

Arguments

dist1	distanct matrix 1
dist2	distanct matrix 2

Value

numeric score

---

Score matrix distances in multiple combinations

Description

Score matrix distances in multiple combinations

Usage

matrix2_score_n(
  dist1,
  dist2,
  n = min(factorial(ncol(dist2)) * 10, 20000),
  verbose = FALSE,
  seed = 34,
  out_worst = FALSE
)

Arguments

dist1	distanct matrix 1
dist2	distanct matrix 2
n	number of iterations
verbose	whether to output more messages
seed	random seed
out_worst	instead of default output of best combination, output worst instead

Value

reordered vector

---