Skip to contents

This function overlays loadings from a PCA or OPLS model on top of NMR spectra using ggplot2. It supports two types of loadings visualization: statistical reconstruction or backscaling.

Usage

specload(
  mod,
  shift = c(0, 10),
  an,
  alp = 0.3,
  size = 0.5,
  pc = 1,
  type = "Backscaled",
  title = "",
  r_scale = FALSE
)

Arguments

mod

A model object of class PCA_metabom8 or OPLS_metabom8.

shift

Numeric vector (length 2), chemical shift region to visualize (in ppm).

an

List of up to three grouping variables (for facet, color, linetype). First must be defined.

alp

Numeric (0–1). Alpha level for spectra.

size

Numeric. Line width for plotted spectra.

pc

Integer. Component to visualize (for OPLS, set to 1).

type

Character. Either "Statistical reconstruction" or "Backscaled" (case-insensitive).

title

Character. Plot title.

r_scale

Logical. If TRUE, correlation color scale is fixed to [0, 1]. Only used in statistical reconstruction.

Value

A ggplot object.

Details

  • Statistical reconstruction: Correlates predictive scores with spectral variables.

  • Backscaled: Multiplies model loadings by feature SDs.

See also

Examples

data(covid)
X <- covid$X
an <- covid$an

model <- pca(X)
specload(model, shift = c(1, 2), an = list(an$type), pc = 1, alp = 0.8)