Profile Line Plot for Heatmap of SummarizedExperiment data

Convert Heatmap of SummarizedExperiment data into Profile Line Plot, with row and column groups and facets as relevant.

Usage

heatmap_profile_plot(
  hm,
  display_profiles = c("rows", "group"),
  summarize_column_split = TRUE,
  se = NULL,
  rowData_colnames = NULL,
  profile_color = "#88888844",
  mean_color = "#000000FF",
  include_count = TRUE,
  row_type = "Gene",
  strip.position = "right",
  base_size = 10,
  panel.background.fill = "grey97",
  facet_ncol = 1,
  return_type = c("ggplot", "list"),
  ...
)

Arguments

hm

ComplexHeatmap::Heatmap object, intended to use output from heatmap_se(). It uses the data within the heatmap to produce the plot, therefore all data centering is used as defined in heatmap_se().

display_profiles

character indicating which profiles to include:

'rows': include a profile line for each row
'group': include a profile line for the row group.

summarize_column_split

logical default TRUE, whether to summarize values in column_split by taking the mean across replicates (columns) within each column grouping.

When summarize_column_split=FALSE it will keep replicates as individual points, which may appear more consistent with the heatmap.

rowData_colnames

character vector, currently ignored. In future it may be used to apply different row groupings.

profile_color

character string, default is partially transparent, light grey. This color is used for the line profile color.

include_count

logical default TRUE, whether to include the row count in the label for each row group.

It uses sep="\n" by default between the row group label, and the row count label. So to make the label all appear on one line, use sep=" ".

row_type

character string, default "Gene" used when include_count=TRUE, so the label becomes "11 Genes" The label should be singular, and the 's' is added when there is more than 1 row in the group.

strip.position

character string, default 'right', places the ggplot2 facet strip label on the right side. The other common option is 'top', which is better for long labels, but requires a taller graphics device to accomodate the number of row groups.

base_size

numeric base default font size for the ggplot2 theme, default 10 is relatively small, suitable for analysis. Use a larger number for figures, but adjust the graphics device size accordingly.

panel.background.fill

character color used for the ggplot2 panel background color, default is very light grey (lighter than ggplot2 default). The grey color makes it slightly easier to recognize each plot panel.

...

additional arguments are passed to internal functions, specifically colorjam::theme_jam() which defines the ggplot2 theme. The theme can be replaced by adding ggplot2::theme_bw() or something else for example.

Value

ggplot object suitable for plotting or further customization.

When return_type='list' it returns a list with 'hmtall' which is a data.frame containing data used for the ggplot. If display_profiles includes 'group' the list will contain 'hmtall4' which is a data.frame containing the row group mean values.
When return_type='ggplot' (default) the output@data can also be used to inspect the underlying data.

Details

It uses the data within the heatmap to produce the plot, therefore all data centering is used as defined in heatmap_se(). This plot is intended to produce a line plot equivalent to the same data shown in heatmap form. Sometimes the linear plot is more visually intuitive, and certainly is more effective at conveying magnitudes than color gradient.

Row groups are detected by calling jamba::heatmap_row_order(). If there are no row groups, all rows are placed into the same group.

Column groups are also detected by calling jamba::heatmap_column_order(). The argument summarize_column_split=TRUE will optionally calculate group mean per column group, but the default is FALSE.

Examples

se <- make_se_test()
hm1 <- heatmap_se(se,
   rowData_colnames="Class",
   row_split=9,
   column_split="group",
   show_left_annotation_name="top",
   left_annotation_name_rot=150,
   sample_color_list=list(group=c(groupA="red")))
#> 'magick' package is suggested to install to give better rasterization.
#> 
#> Set `ht_opt$message = FALSE` to turn off this message.
hm1drawn <- ComplexHeatmap::draw(hm1)

heatmap_profile_plot(hm1drawn, summarize_column_split=FALSE);


se <- jamses::make_se_test(nrow=200, ngroups=4, nreps=8)
hm <- jamses::heatmap_se(se,
   rowData_colnames="Class",
   apply_hm_column_title=TRUE,
   column_split="group",
   cluster_row_slices=TRUE,
   controlSamples=head(colnames(se), 8),
   row_split=12)
#> 'magick' package is suggested to install to give better rasterization.
#> 
#> Set `ht_opt$message = FALSE` to turn off this message.
ComplexHeatmap::draw(hm)


heatmap_profile_plot(hm, strip.position="top")
#> Warning: The heatmap has not been initialized. You might have different results
#> if you repeatedly execute this function, e.g. when row_km/column_km was
#> set. It is more suggested to do as `ht = draw(ht); row_order(ht)`.
#> Warning: The heatmap has not been initialized. You might have different results
#> if you repeatedly execute this function, e.g. when row_km/column_km was
#> set. It is more suggested to do as `ht = draw(ht); column_order(ht)`.
#> 
#> Attaching package: ‘dplyr’
#> The following object is masked from ‘package:jamses’:
#> 
#>     groups
#> The following objects are masked from ‘package:stats’:
#> 
#>     filter, lag
#> The following objects are masked from ‘package:base’:
#> 
#>     intersect, setdiff, setequal, union


heatmap_profile_plot(hm, summarize_column_split=FALSE, strip.position="top")
#> Warning: The heatmap has not been initialized. You might have different results
#> if you repeatedly execute this function, e.g. when row_km/column_km was
#> set. It is more suggested to do as `ht = draw(ht); row_order(ht)`.
#> Warning: The heatmap has not been initialized. You might have different results
#> if you repeatedly execute this function, e.g. when row_km/column_km was
#> set. It is more suggested to do as `ht = draw(ht); column_order(ht)`.


heatmap_profile_plot(hm, summarize_column_split=FALSE)
#> Warning: The heatmap has not been initialized. You might have different results
#> if you repeatedly execute this function, e.g. when row_km/column_km was
#> set. It is more suggested to do as `ht = draw(ht); row_order(ht)`.
#> Warning: The heatmap has not been initialized. You might have different results
#> if you repeatedly execute this function, e.g. when row_km/column_km was
#> set. It is more suggested to do as `ht = draw(ht); column_order(ht)`.