Skip to contents

Plot Point-Biserial Correlations Between Binary and Continuous Variables

Source: R/PlotPointCorrelations.R
PlotPointCorrelationsHeatmap.Rd

Calculates point-biserial correlations (binary vs continuous) with explicit 0/1 coding where 1 == PositiveLevel from createBinaryMapping(), and renders heatmap-style tiles.

Usage

PlotPointCorrelationsHeatmap(
  data,
  CatVars,
  ContVars,
  covariates = NULL,
  Relabel = TRUE,
  Ordinal = TRUE,
  binary_map = NULL,
  fdr_scope = c("matrix", "per_outcome", "per_predictor"),
  Data = lifecycle::deprecated(),
  Covariates = lifecycle::deprecated()
)

Arguments

data

A dataframe.

CatVars

Character vector of binary categorical variables.

ContVars

Character vector of continuous variables.

covariates

Optional covariates (reserved).

Relabel

Logical; use sjlabelled variable labels for axes.

Ordinal

Logical; reserved for future use.

binary_map

Optional mapping as returned by createBinaryMapping(). If NULL, a mapping is created internally for CatVars.

fdr_scope

Either "matrix" (default) or "per_outcome", passed to ApplyFDRCorrection(). "matrix" corrects across all p-values at once (historical behavior). "per_outcome" corrects separately within each continuous variable: outcomes are the continuous variables (ContVars).

Data

Deprecated (since 19.15.0). Use data instead.

Covariates

Deprecated (since 19.15.0). Use covariates instead.

Value

A list with Unadjusted, FDRCorrected, method ("R_pb"), Relabel, Covariates, BinaryMapping.

Examples

data(SampleData)
data(SampleVariableTypes)

Labelled <- RevalueData(SampleData, SampleVariableTypes)$RevaluedData

# CatVars must be binary (exactly two unique non-NA values). These markers
# separate on Diagnosis (Ab_42, tau, p_tau) or on sex (Leptin, PAI_1,
# NT_proBNP), so both rows of the heatmap carry signal in both directions.
result <- PlotPointCorrelationsHeatmap(
  Labelled,
  CatVars = c("Diagnosis", "sex"),
  ContVars = c("Ab_42", "tau", "p_tau", "Leptin", "PAI_1", "NT_proBNP")
)

# Raw p-value point-biserial heatmap
result$Unadjusted$plot
#> Warning: Removed 4 rows containing missing values or values outside the scale range
#> (`geom_text()`).


# FDR-adjusted point-biserial heatmap
result$FDRCorrected$plot
#> Warning: Removed 4 rows containing missing values or values outside the scale range
#> (`geom_text()`).