Skip to contents

Plot Interaction Effects Matrix

Source: R/PlotInteractionEffectMatrix.R
PlotInteractionEffectsMatrix.Rd

Creates a heatmap visualization of interaction effects between continuous variables, showing whether interactions result in slope reversals or maintain the same direction.

Usage

PlotInteractionEffectsMatrix(
  data,
  interVar = NULL,
  outcome_vars = NULL,
  predictor_vars = NULL,
  covariates = NULL,
  Relabel = TRUE,
  Ordinal = lifecycle::deprecated(),
  TreatOrdinalAs = "Exclude",
  Data = lifecycle::deprecated(),
  outcomeVars = lifecycle::deprecated(),
  predictorVars = lifecycle::deprecated(),
  fdr_scope = c("matrix", "per_outcome", "per_predictor"),
  covars = lifecycle::deprecated()
)

Arguments

data

A data frame containing the variables to be analyzed

interVar

Character string specifying the interaction variable (moderator). Can be categorical or continuous.

outcome_vars

Character vector of outcome variable names (displayed on rows)

predictor_vars

Character vector of predictor variable names (displayed on columns)

covariates

Character vector of covariate names to include in the models

Relabel

Logical indicating whether to use variable labels if available (default: TRUE)

Ordinal

Deprecated (since 20.20.0). Use TreatOrdinalAs instead.

TreatOrdinalAs

How ordinal variables are handled. This interaction matrix accepts "Exclude", "Continuous", or "Categorical". Categorical ordinal outcomes are not supported by the linear models.

Data

Deprecated (since 19.15.0). Use data instead.

outcomeVars

Deprecated (since 19.15.0). Use outcome_vars instead.

predictorVars

Deprecated (since 19.15.0). Use predictor_vars instead.

fdr_scope

Either "matrix" (default) or "per_outcome", passed to ApplyFDRCorrection(). "matrix" corrects across all interaction p-values at once (historical behavior). "per_outcome" corrects separately within each outcome variable (outcome_vars).

covars

Deprecated (since 19.15.0). Use covariates instead.

Value

A list containing:

Unadjusted

List with unadjusted results including:

  • C: Matrix of interaction coefficients

  • S: Matrix of slope direction indicators (1 = same, -1 = reversed)

  • P: Matrix of p-values

  • D: Matrix of interaction coefficient signs

  • Slope1: Matrix of slopes at low values of interVar (mean - 1SD for continuous, reference group for categorical)

  • Slope2: Matrix of slopes at high values of interVar (mean + 1SD for continuous, comparison group for categorical)

  • plot: ggplot object of the heatmap

  • pvaltable: P-value table in wide format

FDRCorrected

List with FDR-corrected results (same structure as Unadjusted)

Relabel

Logical indicating whether relabeling was applied

Covariates

Character vector of covariates used

interVar

The interaction variable name

raw_data

Processed data frame with all calculated values

Details

The function fits linear models of the form: outcome ~ predictor * interVar + covariates for each combination of outcome and predictor variables.

For continuous interaction variables, slopes are calculated at mean - 1SD and mean + 1SD. For categorical interaction variables, slopes are calculated for each category.

The function then determines whether the interaction causes a slope reversal (opposite signs) or maintains the same direction (same signs) for the predictor-outcome relationship.

Color coding:

  • Blue gradient: Significant interaction with slopes in the same direction

  • Red gradient: Significant interaction with slope reversal

  • White: Non-significant interaction (p > 0.05)

  • Grey: Missing data or model could not be fit

Darker colors indicate higher significance:

  • ***: p <= 0.001 (darkest)

  • **: p <= 0.01 (medium)

  • *: p <= 0.05 (light)

Examples

# \donttest{
data(SampleData)
data(SampleVariableTypes)

# Attach labels and factor levels for readable axes
Labelled <- RevalueData(SampleData, SampleVariableTypes)$RevaluedData

outcomes <- c("age", "ACE_CD143_Angiotensin_Converti",
              "ACTH_Adrenocorticotropic_Hormon", "AXL", "Adiponectin",
              "Alpha_1_Antichymotrypsin", "Alpha_1_Antitrypsin",
              "Alpha_1_Microglobulin")
predictors <- c("Alpha_2_Macroglobulin", "Angiopoietin_2_ANG_2",
                "Apolipoprotein_A_IV", "Apolipoprotein_A1",
                "Apolipoprotein_A2", "Apolipoprotein_B", "Apolipoprotein_CI",
                "Apolipoprotein_CIII", "Apolipoprotein_D", "Apolipoprotein_E")

# With a categorical interaction variable (Diagnosis)
results <- PlotInteractionEffectsMatrix(
  data = Labelled,
  interVar = "Diagnosis",
  outcome_vars = outcomes,
  predictor_vars = predictors
)

# With a continuous interaction variable (age)
results_cont <- PlotInteractionEffectsMatrix(
  data = Labelled,
  interVar = "age",
  outcome_vars = setdiff(outcomes, "age"),
  predictor_vars = predictors
)

# Raw p-value interaction matrix
results$Unadjusted$plot


# FDR-adjusted interaction matrix
results$FDRCorrected$plot

# }