plot_correlations_distributions: Produce a density plot of correlation values for each window...

In scFeatureFilter: A correlation-based method for quality filtering of single-cell RNAseq data

Description

Feature by feature correlation values between every windows and the reference to window of features are visualized as density lines, one facet per comparison. Two density lines are drown in each facets:

• A thin colored line, the correlations between the bin and the reference top bin of features

• A thicker blue line with grey error area, the correlations between the bin and the randomized top bin of features. The lines are not shown if n_random = 0 in correlate_windows.

Usage

plot_correlations_distributions(df, metrics = NULL, vlines = c("mean",
  "median"), facet_ncol = 4)

Arguments

df	A tibble, usually the output of correlations_to_densities.
metrics	Optional. The output of get_mean_median. Dashed line will represent mean or median of the correlation coefficient distributions.
vlines	A string, either "mean" or "median". Should the dashed line represent the mean or the median of the correlation coefficient distributions? Ignored if metrics is NULL.
facet_ncol	The number of columns to arrange the plots.

Value

A ggplot2 plot.

See Also

correlations_to_densities, get_mean_median

Examples

library(magrittr)
myData <- scData_hESC %>%
calculate_cvs %>%
    define_top_genes(window_size = 100) %>%
    bin_scdata(window_size = 1000)

corDistrib <- correlate_windows(myData, n_random = 3)

corDens <- correlations_to_densities(corDistrib)

plot_correlations_distributions(corDens)

metrics <- get_mean_median(corDistrib)

plot_correlations_distributions(corDens, metrics = metrics)

scFeatureFilter documentation built on Nov. 8, 2020, 7:49 p.m.