Correlation of Principal Components with Sample Variables

This correlate() method analyses the correlations (in R2 values) between all sample variables in a AnyHermesData object and the principal components of the samples.

A corresponding autoplot() method then can visualize the results in a heatmap.

Usage

# S4 method for HermesDataPca
correlate(object, data)

# S4 method for HermesDataPcaCor
autoplot(
  object,
  cor_colors = circlize::colorRamp2(c(-1, 0, 1), c("blue", "white", "red")),
  ...
)

Arguments

object

(HermesDataPca)
input. It can be generated using calc_pca() function on AnyHermesData.

data

(AnyHermesData)
input that was used originally for the PCA.

cor_colors

(function)
color scale function for the correlation values in the heatmap, produced by circlize::colorRamp2().

...

other arguments to be passed to ComplexHeatmap::Heatmap().

Value

A HermesDataPcaCor object with R2 values for all sample variables.

Functions

• autoplot(HermesDataPcaCor): This plot method uses the ComplexHeatmap::Heatmap() function to visualize a HermesDataPcaCor object.

See also

h_pca_df_r2_matrix() which is used internally for the details.

Examples

object <- hermes_data %>%
  add_quality_flags() %>%
  filter() %>%
  normalize()

# Perform PCA and then correlate the prinicipal components with the sample variables.
object_pca <- calc_pca(object)
result <- correlate(object_pca, object)

# Visualize the correlations in a heatmap.
autoplot(result)


# We can also choose to not reorder the columns.
autoplot(result, cluster_columns = FALSE)


# We can also choose break-points for color customization.
autoplot(
  result,
  cor_colors = circlize::colorRamp2(
    c(-0.5, -0.25, 0, 0.25, 0.5, 0.75, 1),
    c("blue", "green", "purple", "yellow", "orange", "red", "brown")
  )
)