Step: Correlation Analysis

Perform pairwise correlation analysis using glystats::gly_cor() and visualize the correlation matrix using glyvis::plot_corrplot(). This step calculates correlation coefficients and p-values for all pairs of variables or samples.

Usage

step_correlation(
  on = "exp",
  on_cor = c("variable", "sample"),
  method = c("pearson", "spearman"),
  p_adj_method = "BH",
  plot_width = 7,
  plot_height = 7,
  ...
)

Arguments

on: Name of the experiment to run correlation analysis on. Can be "exp", "sig_exp", "trait_exp", "sig_trait_exp", "motif_exp", "sig_motif_exp".
on_cor: A character string specifying what to correlate. Either "variable" (default) to correlate variables/features, or "sample" to correlate samples.
method: A character string indicating which correlation coefficient is to be computed. One of "pearson" (default) or "spearman".
p_adj_method: A character string specifying the method to adjust p-values. See p.adjust.methods for available methods. Default is "BH". If NULL, no adjustment is performed.
plot_width: Width of the plot in inches. Default is 7.
plot_height: Height of the plot in inches. Default is 7.
...: Additional arguments passed to glystats::gly_cor().

Value

A glysmith_step object.

Details

Data required:

exp (if on = "exp"): The experiment to run correlation analysis on
trait_exp (if on = "trait_exp"): The trait experiment to run correlation analysis on
motif_exp (if on = "motif_exp"): The motif experiment to run correlation analysis on

Tables generated (with suffixes):

correlation: A table containing pairwise correlation results with columns:
- variable1, variable2 (or sample1, sample2 if on = "sample")
- cor: Correlation coefficient
- p_val: P-value from correlation test
- p_adj: Adjusted p-value (if p_adj_method is not NULL)

Plots generated (with suffixes):

correlation: A correlation matrix heatmap

AI Prompt

This section is for AI in inquire_blueprint() only.

Include this step to explore relationships between variables or samples.
Be careful to use when sample size or variable number is large (> 50). Before using this step for large data, ask the user if they want to proceed.

Examples

step_correlation()
#> <step "step_correlation()"> Correlation analysis
step_correlation(on = "sig_exp")
#> <step "step_correlation(on = \"sig_exp\")"> Correlation analysis of significant
#> variables
step_correlation(on_cor = "sample", method = "spearman")
#> <step "step_correlation(on_cor = \"sample\", method = \"spearman\")">
#> Correlation analysis