Skip to contents

Plot PCA loadings

Source: R/plots-qc-pca.R
plot_pca_loading.Rd

Generates a plot of PCA loadings, illustrating the contribution of features to each principal component. This function can be used to investigate which feature (groups) are contributing to the variance seen in the plot and which need further investigation.

Usage

plot_pca_loading(
  data = NULL,
  variable,
  qc_types = NA,
  pca_dims = c(1, 2, 3, 4),
  log_transform = TRUE,
  top_n = 30,
  vertical_bars = FALSE,
  abs_loading = TRUE,
  filter_data = FALSE,
  include_qualifier = FALSE,
  include_istd = FALSE,
  include_feature_filter = NA,
  exclude_feature_filter = NA,
  min_median_value = NA,
  font_base_size = 8
)

Arguments

data

A MRMhubExperiment object.

variable

A character string indicating the variable to use for PCA analysis. Must be one of: "area", "height", "intensity", "norm_intensity", "response", "conc", "conc_raw", "rt", "fwhm".

qc_types

A character vector specifying the QC types to plot. It must contain at least one element. The default NA plots any of the non-blank QC types ("SPL", "TQC", "BQC", "HQC", "MQC", "LQC", "NIST", "LTR") present in the dataset.

pca_dims

A numeric vector indicating for which PCA dimensions to the loadings should be shown. Default is c(1, 2, 3, 4).

log_transform

A logical value indicating whether to log-transform the data before the PCA. Default is TRUE.

top_n

Number of top features with highest absolute loading that will be to shown for each PC dimension. Default is 30.

vertical_bars

Show vertical bars instead of horizontal bars in the plot. Default is FALSE.

abs_loading

Show absolute loading values instead of signed loadings. Default is TRUE.

filter_data

A logical value indicating whether to use all data (default) or only QC-filtered data (filtered via filter_features_qc()).

include_qualifier

A logical value indicating whether to include qualifier features. Default is TRUE.

include_istd

A logical value indicating whether to include internal standard (ISTD) features. Default is TRUE.

include_feature_filter

A regex pattern or a vector of feature names used to filter features by feature_id. If NA or an empty string ("") is provided, the filter is ignored. When a vector of length > 1 is supplied, only features with exactly these names are selected (applied individually as OR conditions).

exclude_feature_filter

A regex pattern or a vector of feature names used to exclude features by feature_id. If NA or an empty string ("") is provided, the filter is ignored. When a vector of length > 1 is supplied, only features with exactly these names are excluded (applied individually as OR conditions).

min_median_value

Minimum median feature value across the selected QC-type samples required for a feature to be included. NA (default) applies no filtering. This is a fast way to exclude noisy features; for principled QC-based filtering use filter_features_qc().

font_base_size

Numeric. Base font size (in points) for plot text. Default is 8.

Value

A ggplot object with PCA loadings plot

See also

Other QC plots: plot_feature_correlations(), plot_normalization_qc(), plot_pca(), plot_qc_interferences(), plot_qc_matrixeffects(), plot_qc_summary_byclass(), plot_qc_summary_overall(), plot_qcmetrics_comparison(), plot_rla_boxplot(), plot_rt_vs_chain(), plot_runscatter(), plot_runsequence()