Visualize associations between features and classes

This function visualizes different measures of association between features and the label, computed previously with the check.associations function

association.plot(siamcat, fn.plot=NULL, color.scheme = "RdYlBu",
sort.by = "fc", max.show = 50, plot.type = "quantile.box",
panels = c("fc", "auroc"), prompt=TRUE, verbose = 1)

Arguments

siamcat: object of class siamcat-class
fn.plot: string, filename for the pdf-plot. If fn.plot is NULL, the plot will be produced in the active graphics device.
color.scheme: valid R color scheme or vector of valid R colors (must be of the same length as the number of classes), defaults to 'RdYlBu'
sort.by: string, sort features by p-value ("p.val"), by fold change ("fc") or by prevalence shift ("pr.shift"), defaults to "fc"
max.show: integer, how many associated features should be shown, defaults to 50
plot.type: string, specify how the abundance should be plotted, must be one of these: c("bean", "box", "quantile.box", "quantile.rect"), defaults to "quantile.box"
panels: vector, name of the panels to be plotted next to the abundances, possible entries are c("fc", "auroc", "prevalence"), defaults to c("fc", "auroc")
prompt: boolean, turn on/off prompting user input when not plotting into a pdf-file, defaults to TRUE
verbose: integer, control output: 0 for no output at all, 1 for only information about progress and success, 2 for normal level of information and 3 for full debug information, defaults to 1

Value

Does not return anything, but instead produces association plot

Details

This function visualizes the results of the computations carried out in the check.associations function. It produces a plot of the top max.show associated features at a user-specified significance level alpha.

For binary classification problems, the plot will show the distribution of the log10-transformed abundances for both classes, a P-value from the significance test, and user-selected panels for the effect size (AU-ROC, prevalence shift, or generalized fold change). For regression problems, the plot will show the Spearman correlation, the significance, and the linear model effect size.

Examples

# Example data
data(siamcat_example)

# Simple example
association.plot(siamcat_example, fn.plot = "./assoc_plot.pdf")
#> Plotted associations between features and label successfully to: ./assoc_plot.pdf

# Plot associations as box plot
association.plot(siamcat_example,
    fn.plot = "./assoc_plot_box.pdf",
    plot.type = "box")
#> Plotted associations between features and label successfully to: ./assoc_plot_box.pdf

# Additionally, sort by p-value instead of by fold change
association.plot(siamcat_example,
    fn.plot = "./assoc_plot_fc.pdf",
    plot.type = "box", sort.by = "p.val")
#> Plotted associations between features and label successfully to: ./assoc_plot_fc.pdf

# Custom colors
association.plot(siamcat_example,
    fn.plot = "./assoc_plot_blue_yellow.pdf",
    plot.type = "box", color.scheme = c("cornflowerblue", "#ffc125"))
#> Plotted associations between features and label successfully to: ./assoc_plot_blue_yellow.pdf