mbecPCA: Principal Component Analysis Plot

In buschlab/MBECS: Evaluation and correction of batch effects in microbiome data-sets

Description

Takes two covariates, i.e., group and batch, and computes the ordination-plot for user-selected principal components. Covariates determine sample-shape and color and can be switched to shift the emphasis on either group. In addition to the ordination-plot, the function will show the distribution of eigenvalues (colored by the second covariate) on their respective principal components.

Usage

mbecPCA(
  input.obj,
  model.vars = c("batch", "group"),
  pca.axes = c(1, 2),
  type = "clr",
  label = character(),
  return.data = FALSE
)

Arguments

input.obj	list(cnts, meta), phyloseq, MbecData object (correct orientation is handled internally)
model.vars	two covariates of interest to select by first variable selects color (batch) and second one determines shape (group)
pca.axes	numeric vector which axes to plot, first is X and second is Y
type	Which abundance matrix to use for the calculation.
label	Which corrected abundance matrix to use for analysis.
return.data	logical if TRUE returns the data.frame required for plotting. Default (FALSE) will return plot object.

Details

The function returns either a plot-frame or the finished ggplot object. Input is an MbecData-object. If cumulative log-ratio (clr) and total sum-scaled (tss) abundance matrices are part of the input, i.e., 'mbecTransform()' was used, they can be selected as input by using the 'type' argument with either "otu", "clr" or "tss". If batch effect corrected matrices are available, they can be used by specifying the 'type' argument as "cor" and using the 'label' argument to select the appropriate matrix by its denominator, e.g., for batch correction method ComBat this would be "bat", for RemoveBatchEffects from the limma package this is "rbe". Default correction method-labels are "ruv3", "bmc","bat","rbe","pn","svd".

The combination of 'type' and 'label' argument basically accesses the attribute 'cor', a list that stores all matrices of corrected counts. This list can also be accessed via getter and setter methods. Hence, the user can supply their own matrices with own names.

Value

either a ggplot2 object or a formatted data-frame to plot from

Examples

# This will return the data.frame for plotting.
data.PCA <- mbecPCA(input.obj=dummy.mbec,
model.vars=c('group','batch'), pca.axes=c(1,2), return.data=TRUE)

# This will return the ggplot2 object for display, saving and modification.
# Selected PCs are PC3 on x-axis and PC2 on y-axis.
plot.PCA <- mbecPCA(input.obj=dummy.mbec,
model.vars=c('group','batch'), pca.axes=c(3,2), return.data=FALSE)

buschlab/MBECS documentation built on Jan. 21, 2022, 1:27 a.m.