Skip to contents

Experimental prolong Functions

Source: vignettes/prolong-experimental.Rmd
prolong-experimental.Rmd
library(prolong)
data("sim_outcome")
data("sim_metabs")

Warning

These functions can produce additional visualizations but are not as soundly backed by our methodology and are subject to drastic change throughout versions of prolong.

delta_scatter() Function

The delta_scatter() function can show a 2D or 3D scatterplot of pairwise delta-scale correlations to view the relationship between these pairwise correlations across time. The default is a 2D interactive scatterplot with hovertext for each point, and with fisher z-transformation applied to the pairwise correlations.

delta_scatter(sim_metabs)

If interactive = FALSE a non-interactive version will be shown without the hovertext. We can also change timediff1 and timediff2 to compare different time point pairs. If we set fisherz = FALSE, the raw pairwise correlations will be shown.

delta_scatter(sim_metabs, timediff1 = "3-2", timediff2 = '4-3', interactive = FALSE, fisherz = FALSE)

For 3D scatterplots, we can repeat the above but specify timediff3. Setting interactive = FALSE will not change anything if timediff3 is supplied, only an interactive option is currently available. We use a subset of the variables here to reduce the file size of the rendered html.

delta_scatter(sim_metabs[,1:20,], timediff3 = '4-3')

delta_network() Function

The delta_network() function plots a network where the nodes are the \(p\) variables and edges exist only between pairs of variables whose (partial) pearson correlation is greater than corr_thresh. The default is an interative network plot with edges between pairs of variables with partial correlation greater than \(0.75\).

delta_network(sim_metabs)

As in delta_scatter() we can view a static plot instead with interactive = FALSE. We can also change the correlation threshold with corr_thresh, set partial = FALSE to use regular pairwise correlations, and change method to either "kendall" or "spearman" instead of the default "pearson".

delta_network(sim_metabs, timediff = "3-2", partial = FALSE, corr_thresh = 0.8, interactive = FALSE, method = "spearman")