ggplate

CRAN status R-CMD-check Codecov test coverage

The goal of the ggplate package is to enable users to create simple plots of biological culture plates as well as microplates. Both continuous and discrete values can be plotted onto the plate layout.

Currently the package supports the following plate sizes:

Installation

ggplate is implemented as an R package.

You can install the release version from CRAN using the install.packages() function.

install.packages("ggplate")

You can install the development version from GitHub using the devtools package by copying the following commands into R:

Note: If you do not have devtools installed make sure to do so by removing the comment sign (#).

# install.packages("devtools")
devtools::install_github("jpquast/ggplate")

In addition, you can install it via the command line through conda since it is also implemented as a conda-forge package.

conda install -c conda-forge r-ggplate

Usage

In order to use ggplate you have to load the package in your R environment by simply calling the library() function as shown bellow.

# Load ggplate package
library(ggplate)

There are multiple example datasets provided that can be used to create plots of each plate type. You can access these datasets using the data() function.

# Load a dataset of continuous values for a 96-well plate
data(data_continuous_96)

# Check the structure of the dataset
str(data_continuous_96)
#> Classes 'tbl_df', 'tbl' and 'data.frame':    96 obs. of  2 variables:
#>  $ Value: num  1.19 0.88 0.17 0.85 0.78 0.23 1.95 0.4 0.88 0.26 ...
#>  $ well : chr  "A1" "A2" "A3" "A4" ...

When calling the str() function you can see that the data frame of a continuous 96-well plate dataset only contains two columns. The Value column contains values associated with each of the plate wells, while the well column contains the corresponding well positions using a combination of alphabetic row names and numeric column names.

You can use this example data frame to create a 96-well plate layout plot using the plate_plot() function and setting the plate_size argument to 96. There are currently two options for the plate well type. These can be either "round" or "square". In the plot below we specify "round", while "square" is the default value when the plate_type argument is not provided.

The data frame is provided to the data argument and the column name of the column containing the well positions is provided to the position argument. The column name of the column containing the values is provided to the value argument.

Note: For an R markdown file set the chunk options to dpi=300 for an optimal result.

# Create a 96-well plot with round wells
plate_plot(
  data = data_continuous_96,
  position = well,
  value = Value,
  plate_size = 96,
  plate_type = "round"
)

It is also possible to label each well in the plate with a corresponding label. For the plate above it would be interesting to display the exact value on each of the wells in addition to the colouring. For that we use the label argument which takes the name of the column containing the label as an input. In this example case this column is the same that is also provided to the value argument.

# Create a 96-well plot with labels
plate_plot(
  data = data_continuous_96,
  position = well,
  value = Value,
  label = Value,
  plate_size = 96,
  plate_type = "round"
)

Try providing the well column to the label argument instead of the Value column. This will label each will with its position, which might make it easier to find specific positions.

# Create a 96-well plot with labels
plate_plot(
  data = data_continuous_96,
  position = well,
  value = Value,
  label = well,
  plate_size = 96,
  plate_type = "round"
)

Legend Limit Adjustment

The legend for continuous values will only cover a range from the minimal measured to the maximal measured value. If the theoretically expected range of values is however bigger than the measured range you can adjust the legend limits. This can be done using the limits arguments. You provide a vector with the new minimum and maximum to the argument. Use NA to refer to the existing minimum or maximum if you only want to adjust one. Below we show this for an example dataset of a 384-well plate.

# Load a dataset of continuous values for a 384-well plate
data(data_continuous_384)

# Check the structure of the dataset
str(data_continuous_384)
#> tibble [384 × 2] (S3: tbl_df/tbl/data.frame)
#>  $ Value: num [1:384] 1.98 0.7 0.61 0.12 0.27 0.67 0.47 0.37 0.04 0.65 ...
#>  $ well : chr [1:384] "A1" "A2" "A3" "A4" ...

# Create a 384-well plot with adjusted legend limits
plate_plot(
  data = data_continuous_384,
  position = well,
  value = Value,
  plate_size = 384,
  limits = c(0, 4)
)

If your new range will be smaller than the measured range, values outside of the range are coloured gray.

# Create a 384-well plot with adjusted legend limits and outliers
plate_plot(
  data = data_continuous_384,
  position = well,
  value = Value,
  plate_size = 384,
  limits = c(0, 3)
)

Gradient Colour Adjustment

When plotting continuous variables it is possible to to change the gradient colours by providing new colours to the colour argument. The colours will be used to create a new colour gradient for the plot.

# Create a 384-well plot with a new colour gradient
plate_plot(
  data = data_continuous_384,
  position = well,
  value = Value,
  plate_size = 384,
  colour = c(
    "#000004FF",
    "#51127CFF",
    "#B63679FF",
    "#FB8861FF",
    "#FCFDBFFF"
  )
)

Incomplete datasets

If you have a dataset that does not contain a value for each well, empty wells will be uncoloured. Empty wells can either contain NA as their value argument or they can be completely omitted from the input data frame.

# Load a continuous of discrete values for a 48-well plate
data(data_continuous_48_incomplete)

# Check the structure of the dataset
str(data_continuous_48_incomplete)
#> tibble [48 × 2] (S3: tbl_df/tbl/data.frame)
#>  $ Value: num [1:48] 1.14 0.46 0.72 0.17 NA NA NA NA 1.37 0.37 ...
#>  $ well : chr [1:48] "A1" "A2" "A3" "A4" ...

# Create a 48-well plot with adjusted legend limits
plate_plot(
  data = data_continuous_48_incomplete,
  position = well,
  value = Value,
  plate_type = "round",
  plate_size = 48
)

Plot Customisation

You can further customise your plot in various ways. Lets take a discrete 6-well plate dataset as an example. This dataset only contains three categories assigned to the six wells of the plate. This could be for example a pipetting scheme of an experiment.

You can change the title of the plot using the title argument. In addition the size of the title can be adjusted using the title_size argument.

Note: Using the R markdown chunk options out.width and fig.align you can reduce the size of the figure in the R markdown document and align it for example to the center.

# Load a dataset of discrete values for a 6-well plate
data(data_discrete_6)

# Check the structure of the dataset
str(data_discrete_6)
#> tibble [6 × 2] (S3: tbl_df/tbl/data.frame)
#>  $ Condition: chr [1:6] "DMSO" "Rapamycin" "Taxol" "DMSO" ...
#>  $ well     : chr [1:6] "A1" "A2" "A3" "B1" ...

# Create a 6-well plot with new title
plate_plot(
  data = data_discrete_6,
  position = well,
  value = Condition,
  plate_size = 6,
  plate_type = "round",
  title = "Drug Treatment",
  title_size = 23
)

In addition it is possible to change the colours of the plot by providing new colours to the colour argument. As mentioned earlier this does not only work for discrete values but also for gradients that will be created based on the provided colours.

# Create a 6-well plot
plate_plot(
  data = data_discrete_6,
  position = well,
  value = Condition,
  plate_size = 6,
  plate_type = "round",
  title = "Drug Treatment",
  title_size = 23,
  colour = c("#3a1c71", "#d76d77", "#ffaf7b")
)

Also for this plot we can provide a column name to the label argument to directly label the wells in the plot. At the same time we can disable the legend setting the show_legend argument to FALSE. In this case the labels for each well are too large and we should also resize the label so that it fits perfectly into each well using the label_size argument.

# Create a 6-well plot
plate_plot(
  data = data_discrete_6,
  position = well,
  value = Condition,
  label = Condition,
  plate_size = 6,
  plate_type = "round",
  title = "Drug Treatment",
  title_size = 23,
  colour = c("#3a1c71", "#d76d77", "#ffaf7b"),
  show_legend = FALSE,
  label_size = 4
)

Potential Issues

In order to have the same proportions independent on the output screen and size, each plate plot is scaled according to the specific graphics device size. In order to see the currently used graphics device size and scaling factor the silent argument of the function can be set to FALSE.

As you can see for the bellow example the graphics device size is width: 7 height: 4 and the scaling factor is 1.256.

# Load a dataset of discrete values for a 24-well plate
data(data_discrete_24)

# Check the structure of the dataset
str(data_discrete_24)
#> tibble [24 × 2] (S3: tbl_df/tbl/data.frame)
#>  $ Condition: chr [1:24] "siControl" "siRaptor" "siRagB" "siRagD" ...
#>  $ well     : chr [1:24] "A1" "A2" "A3" "A4" ...

# Create a 24-well plot
plate_plot(
  data = data_discrete_24,
  position = well,
  value = Condition,
  plate_size = 24,
  plate_type = "round",
  silent = FALSE
)
#> width: 7 height: 4
#> scaling factor: 1.256

It is possible that the generated plot has overlapping or too spaced out wells. This can be corrected by resizing the output graphics device size until the plot has the desired proportions. If a specific output size is required and the plot does not have the desired proportions you can use the scale argument to adjust it as shown below.

Note: If you run the package directly in the command line, the function opens a new graphics device since it is not already opened like it would be the case in RStudio. If this is not desired you can avoid this by setting the scale argument.

# Create a 24-well plot
plate_plot(
  data = data_discrete_24,
  position = well,
  value = Condition,
  plate_size = 24,
  plate_type = "round",
  silent = FALSE,
  scale = 1.45
)
#> width: 7 height: 4
#> scaling factor: 1.45

As you can see, however, now we are running into the problem that the legend is larger than the screen size. With the legend_n_row argument you can manually determine the number of rows that should be used for the legend. In this case it is ideal to split the legend into 2 columns by setting legend_n_row to 6 rows. In addition we should adjust the scale parameter to 1.2 in order to space out wells properly.

# Create a 24-well plot with 2 row legend
plate_plot(
  data = data_discrete_24,
  position = well,
  value = Condition,
  plate_size = 24,
  plate_type = "round",
  silent = FALSE,
  scale = 1.2,
  legend_n_row = 6
)
#> width: 7 height: 4
#> scaling factor: 1.2

If your dataset has a lot of labels it can become difficult to impossible to distinguish them just by colour as you can see for the dataset below.

# Load a dataset of discrete values for a 96-well plate
data(data_discrete_96)

# Check the structure of the dataset
str(data_discrete_96)
#> tibble [96 × 3] (S3: tbl_df/tbl/data.frame)
#>  $ Compound          : chr [1:96] "Positive Control" "Axitinib" "Gefitinib (ZD1839)" "Rapamycin (Sirolimus)" ...
#>  $ well              : chr [1:96] "A1" "A2" "A3" "A4" ...
#>  $ Compound_multiline: chr [1:96] "Positive\nControl" "Axitinib" "Gefitinib\n(ZD1839)" "Rapamycin\n(Sirolimus)" ...

# Create a 96-well plot
plate_plot(
  data = data_discrete_96,
  position = well,
  value = Compound,
  plate_size = 96,
  scale = 0.95,
  plate_type = "round"
)

This is an example where it is likely better to directly label wells instead of displaying a legend.

# Create a 96-well plot with labels
plate_plot(
  data = data_discrete_96,
  position = well,
  value = Compound,
  label = Compound_multiline, # using a column that contains line brakes for labeling
  plate_size = 96,
  show_legend = FALSE, # hiding legend
  label_size = 1.1, # setting label size
  plate_type = "round"
)

Figure Export

Since the plot function checks the size of the graphics device in order to apply the appropriate scaling to the plot, it is important to first generate an output graphics device with the correct size. There are several functions that can accomplish this. These include e.g. png(), pdf(), svg() and many more.

# Generate a new graphics device with a defined size
png("plate_plot_384_well_plate.png", width = 10, height = 6, unit = "in", res = 300)

# Create a plot
plate_plot(
  data = data_continuous_384,
  position = well,
  value = Value,
  label = Value,
  plate_size = 384,
  colour = c(
    "#000004FF",
    "#51127CFF",
    "#B63679FF",
    "#FB8861FF",
    "#FCFDBFFF"
  )
)

# Close graphics device
dev.off()