Skip to contents

Generate a heatmap of normalised gene expression values, z-score or log2Fold-change values.

Source: R/rnaseq_related.R
get_gene_expression_heatmap.Rd

Heatmap is a common tool to show gene expression pattern across samples in RNA-seq experiments. During data exploration, it is a common practice to generate several heatmaps to identify interesting patterns of gene expression across samples. Most heatmap generating tools require data in a tabular format. However, to prepare such a data requires significant data wrangling such as sub-setting genes (rows) and samples (columns) from multiple R objects. Such data wrangling creates redundant, and less readable code which increases the chances of error. Furthermore, because of less readability using the same code for another analysis requires code cleaning and code rewriting making code less reusable. Situations complicates even more when RNA-seq studies consist of several samples and several comparisons. This function cut-downs several steps of data wrangling to create a heatmap of gene expression, z-score or log2 fold-change. Given an object of the class parcutils, sample names or sample comparisons, genes to show in the heatmap and several other arguments, the heatmap can be created quickly. The output heatmap is an output of the function ComplexHeatmap::Heatmap() which then can be used with other functions of the ComplexHeatmap package.

Usage

get_gene_expression_heatmap(
  x,
  samples,
  genes,
  repair_genes = FALSE,
  convert_log2 = FALSE,
  color_default = TRUE,
  col = NULL,
  convert_zscore = TRUE,
  convert_rowmeans = FALSE,
  summarise_replicates = TRUE,
  summarise_method = "median",
  show_row_names = FALSE,
  cluster_rows = TRUE,
  show_row_dend = TRUE,
  row_names_font_size = 10,
  label_specific_rows = NULL,
  label_specific_rows_gp = gpar(fontface = "plain", fontsize = 10, col = "red"),
  show_column_names = TRUE,
  cluster_columns = TRUE,
  show_heatmap_legend = TRUE,
  ...
)

get_fold_change_heatmap(
  x,
  sample_comparisons,
  genes,
  repair_genes = FALSE,
  color_default = TRUE,
  col = NULL,
  show_row_names = FALSE,
  cluster_rows = TRUE,
  show_row_dend = TRUE,
  row_names_font_size = 10,
  show_column_names = TRUE,
  cluster_columns = TRUE,
  show_heatmap_legend = TRUE,
  ...
)

Arguments

x

an abject of the class parcutils. This is can be created using the function run_deseq_analysis().

samples

a character vector denoting sample names to use in the heatmap.

genes

a character vector denoting genes to use in the heatmap.

repair_genes

logical, default FALSE, indicating whether to repair gene names. See details.

convert_log2

logical, default FALSE, indicating whether to log2 transform gene expression values.

color_default

logical, default TRUE, indicating whether to use default heatmap colors.

col

an output of circlize::colorRamp2(), default NULL.

convert_zscore

logical, default TRUE, indicating whether to convert gene expression values in to the z-score or not. see details.

convert_rowmeans

logical, default FALSE, indicating whether to subtract gene expression values from mean of the each row. It is recommended to use convert_log2 = TRUE while using convert_rowmeans = TRUE.

summarise_replicates

logical, default TRUE, indicating whether to summarise values for each gene across replicates.

summarise_method

a character string, either mean or median.

show_row_names

logical, default FALSE, indicating whether to show row names in the heatmap or not. Internally this argument is passed to the function ComplexHeatmap::Heatmap().

cluster_rows

logical, default TRUE, indicating whether to cluster rows in the heatmap or not. Internally this argument is passed to the function ComplexHeatmap::Heatmap().

show_row_dend

logical, default TRUE, indicating whether to show dendrogram in the heatmap or not. Internally this argument is passed to the function ComplexHeatmap::Heatmap().

row_names_font_size

a numeric value, default 10, indicating size of row names in the heatmap.

label_specific_rows

a character string denoting set of gene/s to be labelled in the heatmap. Label must match to the rownames of the final matrix supplied.

label_specific_rows_gp

pass to labels_gp of the function ComplexHeatmap::anno_mark().

show_column_names

logical, default TRUE, indicating whether to show column names in the heatmap or not. Internally this argument is passed to the function ComplexHeatmap::Heatmap().

cluster_columns

logical, default TRUE, indicating whether to cluster columns in the heatmap or not. Internally this argument is passed to the function ComplexHeatmap::Heatmap().

show_heatmap_legend

logical, default TRUE, indicating whether to show heatmap legend or not.

...

Other parameters to be passes to ComplexHeatmap::Heatmap().

sample_comparisons

a character vector denoting sample comparisons for which heatmap of log2 fold-change to be plotted.

Value

an output of the function ComplexHeatmap::Heatmap().

Details

The function get_gene_expression_heatmap is to create a heatmap either for normalised gene expression or z-score values while the function get_fold_change_heatmap is to create a heatmap for log2 fold-changes values.

  • repair_genes : Internally gene names are stored as a "gene_id:gene_symbol" format. For example, "ENSG00000187634:SAMD11". When repair_genes is set to TRUE the string corresponding to gene_id followed by ":" will be removed. This is useful when gene names to be revealed in the heatmap.

  • convert_zscore : When set to TRUE values for each gene is converted into z-score. z-score is calculated by baser r function base::scale() with all default parameters.

Examples


count_file <- system.file("extdata","toy_counts.txt" , package = "parcutils")
count_data <- readr::read_delim(count_file, delim = "\t")
#> Rows: 5000 Columns: 10
#> ── Column specification ────────────────────────────────────────────────────────
#> Delimiter: "\t"
#> chr (1): gene_id
#> dbl (9): control_rep1, control_rep2, control_rep3, treat1_rep1, treat1_rep2,...
#> 
#> ℹ Use `spec()` to retrieve the full column specification for this data.
#> ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

sample_info <- count_data %>% colnames() %>% .[-1]  %>%
 tibble::tibble(samples = . , groups = rep(c("control" ,"treatment1" , "treatment2"), each = 3) )


res <- run_deseq_analysis(counts = count_data ,
                         sample_info = sample_info,
                         column_geneid = "gene_id" ,
                         group_numerator = c("treatment1", "treatment2") ,
                         group_denominator = c("control"))
#> ℹ Running DESeq2 ...
#> converting counts to integer mode
#> Warning: some variables in design formula are characters, converting to factors
#> estimating size factors
#> estimating dispersions
#> gene-wise dispersion estimates
#> mean-dispersion relationship
#> final dispersion estimates
#> fitting model and testing
#> ✔ Done.
#> ℹ Summarizing DEG ...
#> ✔ Done.

genes = parcutils::get_genes_by_regulation(x = res, sample_comparison = "treatment2_VS_control" , "both") %>% names()
get_gene_expression_heatmap(x = res, samples = c("control" ,"treatment1" , "treatment2"), genes = genes, label_specific_rows = c("POU5F1","PTPN18","RABL3","PTPN19"),repair_genes = TRUE,show_row_names = FALSE)
#> ! Label PTPN19 is not found in the data.


# plot raw expression values

get_gene_expression_heatmap(x = res, samples = c("control" ,"treatment1" , "treatment2"),
genes = genes, convert_zscore = FALSE)


# plot log2 expression values

get_gene_expression_heatmap(x = res, samples = c("control" ,"treatment1" , "treatment2"),
genes = genes, convert_zscore = FALSE,convert_log2 = TRUE)


# plot all replicates

get_gene_expression_heatmap(x = res, samples = c("control" ,"treatment1" , "treatment2"),
genes = genes, convert_zscore = TRUE, summarise_replicates = FALSE)


# show gene names

get_gene_expression_heatmap(x = res, samples = c("control" ,"treatment1" , "treatment2"),
genes = genes[1:10], convert_zscore = TRUE, summarise_replicates = FALSE , show_row_names = TRUE)


# repair gene names

get_gene_expression_heatmap(x = res, samples = c("control" ,"treatment1" , "treatment2"),
genes = genes[1:10], convert_zscore = TRUE, summarise_replicates = FALSE, show_row_names = TRUE, repair_genes = TRUE)


# fold change heatmap

get_fold_change_heatmap(x = res , sample_comparisons = c("treatment2_VS_control" ,"treatment1_VS_control") ,
genes = genes , cluster_columns = FALSE , name = "Log2FC")