Visualize PCA or OPLS Loadings for NMR Data

Overlay PCA or OPLS loadings on the ppm axis for NMR data. Visualizations can be based on statistical reconstruction or backscaling, providing insights into spectral regions that contribute most to variation or separation.

Usage

plotload(
  mod,
  shift = c(0, 10),
  pc = 1,
  type = "Backscaled",
  title = NULL,
  r_scale = FALSE
)

Arguments

mod: A PCA or OPLS model object generated via the metabom8 package.
shift: Numeric vector of length 2. Chemical shift (ppm) range to display.
pc: Integer. Index of component to visualize (use 1 for OPLS).
type: Character. Either "Statistical reconstruction" or "Backscaled" (case-insensitive).
title: Optional plot title.
r_scale: Logical. If TRUE, correlation color gradient is fixed between 0 and 1 (only for statistical reconstruction).

Value

A ggplot2 object.

Details

For OPLS:

Statistical reconstruction visualizes correlation (r) and covariance between scores and predictors.
Backscaled multiplies loadings by feature standard deviations, highlighting consistent spectral influence.

For PCA, only statistical reconstruction is available.

References

Cloarec, O., et al. (2005). Anal. Chem. 77(2), 517–526.

Examples

data(covid)
X <- covid$X

model <- pca(X)
plotload(model, pc = 1)