Correlation heatmaps

• plot_cor_heatmap(): lower triangle (diagonal included) of a correlation matrix from cor_matrix_test(), with values and stars, variables optionally ordered by hierarchical clustering.

• plot_cor_cross(): rectangular heatmap of the correlations between two sets of variables (e.g. biomarkers x scores), optionally partial (adjusted for covariates, see cor_partial_spearman()).

Usage

plot_cor_heatmap(
  cm,
  labels = NULL,
  use_adj = TRUE,
  order = c("hclust", "original"),
  show_values = TRUE,
  base_size = 10,
  title = NULL,
  value_size = 2.6
)

plot_cor_cross(
  data,
  x_vars,
  y_vars,
  covariates = NULL,
  labels = NULL,
  p_adjust = "BH",
  base_size = 10,
  title = NULL
)

Arguments

cm

Output of cor_matrix_test().

labels

Optional named vector of labels.

use_adj

Stars on adjusted (TRUE) or raw p-values.

order

"hclust" (Ward clustering on 1 - r) or "original".

show_values

Print the correlations and stars.

base_size

Base font size.

title

Title.

value_size

Size of the printed values.

data

A data frame.

x_vars, y_vars

Variables in rows and columns (names or tidyselect).

covariates

Optional adjustment covariates (names or tidyselect).

p_adjust

Multiplicity correction over all the pairs.

Value

plot_cor_heatmap(): a ggplot. plot_cor_cross(): list(table, plot).

Examples

cm <- cor_matrix_test(cohort_sim, c(NFL_V0, IL6_V0, CRP_V0, MOCA_V0, UPDRS_V0, AGE))
plot_cor_heatmap(cm)

res <- plot_cor_cross(cohort_sim, c(NFL_V0, IL6_V0, CRP_V0), c(MOCA_V0, UPDRS_V0),
                      covariates = c(AGE, SEX))
res$plot