Skip to contents

Plot pretty heatmap for single-agent or combo data to control quality of the data

Source: R/plot-pheatmap.R
pheatmap_qc.Rd

Plot pretty heatmap for single-agent or combo data to control quality of the data

Usage

pheatmap_qc(
  dt_average,
  normalization_type = "GR",
  metric = "x",
  fit_source = "gDR",
  hm_title = NA,
  colors_vec = c("black", "grey99"),
  no_breaks = 50,
  cluster_rows = TRUE,
  distfun = compute_distances,
  lbl_by_CellLineName = FALSE,
  lbl_by_DrugName = FALSE
)

Arguments

dt_average

data.table representing data from the Averaged assay, outputted by gDRutils::convert_se_assay_to_dt and SummarizedExperiment with chosen data type: single-agent or combo

normalization_type

string with normalization types to be selected one of: "GR" ("GRvalue") or "RV" ("RelativeViability")

metric

string name of the metric; one of: "x" (value of "GR" or "RV" itself - respectively depending on normalization_type), or "x_std" (standard deviation)

fit_source

string source name for metrics

hm_title

string plot title

colors_vec

character vector of colors (valid name or hex) used in heatmap note that for metric "x" the first color will be assigned to the min value, and the last one - to the max; for "x_std" - that will be reversed

no_breaks

numeric number of breaks on scale used for mapping values to colors

cluster_rows

logical flag whether rows should be clustered; the dendrogram will not be shown for the matrix with any dimension greater than 200.

distfun

function used to compute the distance (dissimilarity) between rows; used for the dendrogram when cluster_rows is set to TRUE; the default is compute_distances using Spearman method.

lbl_by_CellLineName

logical flag whether heatmap should be described by CellLineNames instead of clid

lbl_by_DrugName

logical flag whether heatmap should be described by DrugName instead of Gnumber

Value

pheatmap object containing heatmap for selected metric with annotation - if given

See also

pheatmap::pheatmap

Examples

mae <- gDRutils::get_synthetic_data("combo_matrix")
se <- mae[[gDRutils::get_supported_experiments("sa")]][2:5, ]
dt_average <- gDRutils::convert_se_assay_to_dt(se = se,
                                               assay_name = "Averaged")

hm_1 <- pheatmap_qc(dt_average = dt_average)

hm_2 <- pheatmap_qc(dt_average = dt_average,
                    normalization_type = "RV",
                    colors_vec = c("darkblue", "grey90"),
                    lbl_by_CellLineName = TRUE,
                    lbl_by_DrugName = TRUE)

patchwork::wrap_elements(hm_1[["gtable"]])

patchwork::wrap_elements(hm_2[["gtable"]])


se <- mae[[gDRutils::get_supported_experiments("combo")]]
dt_average <- gDRutils::convert_se_assay_to_dt(se = se,
                                               assay_name = "Averaged")
hm_3 <- pheatmap_qc(dt_average = dt_average,
                    cluster_rows = FALSE)
hm_4 <- pheatmap_qc(dt_average = dt_average,
                    metric = "x_std",
                    cluster_rows = FALSE)

patchwork::wrap_elements(hm_3[["gtable"]])

patchwork::wrap_elements(hm_4[["gtable"]])