The goal of simpleHM is to provide a package to generate heatmap visualisations. In contrast to something like the ComplexHeatmap package, it generates plots using ggplot2 and the tidyverse, allowing for better customisation and integration into existing workflows.
You can install the development version of simpleHM from GitHub with:
# install.packages("devtools")
devtools::install_github("jaspitzer/simpleHM")This is a basic example which shows you how to solve a common problem:
library(simpleHM)
df <- data.frame(samples = c(paste0("untreated", 1:6), paste0("treated", 7:12)),
group = c(rep("Untreated",6), rep("Treated",6)),
patient = c(rep(paste0("Patient", 1:3), 4)),
batch = rep(1:2, 6),
var1 = c(rnorm(6, 10, 1), rnorm(6, 7, .7)),
var2 = c(rnorm(6, 10, 1), rnorm(6, 200, 10)),
var3 = c(rnorm(6, 50, 5), rnorm(6, 10, 1)),
var4 = c(rnorm(6, 10, 1), rnorm(6, 60, .7)))
head(df)
#> samples group patient batch var1 var2 var3 var4
#> 1 untreated1 Untreated Patient1 1 10.080857 9.031367 44.16176 11.945974
#> 2 untreated2 Untreated Patient2 2 9.652045 9.068599 54.38544 9.530221
#> 3 untreated3 Untreated Patient3 1 11.320619 8.654449 45.62264 9.005435
#> 4 untreated4 Untreated Patient1 2 9.906732 12.187999 50.92283 11.006231
#> 5 untreated5 Untreated Patient2 1 9.854646 10.365214 47.04367 11.655614
#> 6 untreated6 Untreated Patient3 2 10.386753 9.448740 50.60357 8.606055
heatmap_plot <- simpleHM(df, excluded_vars = "batch")
heatmap_plot
By default the sample names for <50 samples or >100 params are hidden.
You can also hide them by using the show_sample_names and
show_param_names arguments.
heatmap_plot_no_sample_names <- simpleHM(df, excluded_vars = "batch", show_sample_names = F)
heatmap_plot_no_sample_names
heatmap_plot_no_param_names <- simpleHM(df, excluded_vars = "batch", show_param_names = F)
heatmap_plot_no_param_names
You can also add Annotation bars for several of the conditions by using
the add_annotation parameter:
hm_with_anno <- simpleHM(df, add_annotation = T, anno_col = c("group", "patient"), excluded_vars = "batch")
hm_with_anno
By default, the samples and parameters are clustered. To display the
dendrograms, set add_dendros=T.
hm_with_dendro <- simpleHM(df, excluded_vars = "batch", add_dendros = T)Dendrograms are a for to visualise hierarchical clustering, but the
branches can be rotated freely, without a change in clustering or
distance between branches. For this purpose, the pull_top and
pull_side arguments can be used to “pull” certain samples to the front
through rotation along the branches of the tree.
library(patchwork)
p1 <- simpleHM(df, excluded_vars = "batch", add_dendros = T, hide_legend = T)
p2 <- simpleHM(df, excluded_vars = "batch", add_dendros = T, pull_top = 7:12, hide_legend = T)
p3 <- simpleHM(df, excluded_vars = "batch", add_dendros = T, pull_side = 3:4, hide_legend = T)
p4 <- simpleHM(df, excluded_vars = "batch", add_dendros = T, pull_side = c("var1", "var3"), hide_legend = T)
combined_plots1 <- wrap_plots(list(p1, p2), ncol = 2)
combined_plots2 <- wrap_plots(list(p3, p4), ncol = 2)
combined_plots1 
combined_plots2
These options can also be combined:
annotated_hm_with_dendro <- simpleHM(df, excluded_vars = "batch", add_annotation = T, anno_col = c("group", "patient"), add_dendros = T)
annotated_hm_with_dendro
You can also generate the separate parts like the heatmap, the dendrogram and the annotation seperately and later combine them in a way you see fit.
just_dendros <- simpleDendro(df, excluded_vars = "batch")
just_dendros
#> $top
#>
#> $side
just_annotations <- simpleAnno(df, excluded_vars = "batch")
just_annotations
#> $batch
#>
#> $patient
#>
#> $group
As a
note, the excluded_vars argument excludes the (numeric) variable from
influence in clustering and normalisation, it does not exclude it from
the annotation columns. You can however specify the columns through the
annotation_cols parameter:
specific_annotations <- simpleAnno(df, anno_col = c("group", "patient"))
specific_annotations
#> $patient
#>
#> $group
