Randomise reaction coordinates

Randomise the coordinates of the reactions in a plate, so that reactions with similar parameters are less likely to be affected by the same spatial technical biases. This is also useful when performing plate replicates. Finally, the coordinates of negative controls will usually make a non-symmetric pattern that helps to detect when a plate was rotated accidentally.

Usage

randomise(plate, seed, vector)

# S4 method for Plate,missing,missing
randomise(plate, seed, vector)

# S4 method for Plate,numeric,missing
randomise(plate, seed, vector)

# S4 method for Plate,missing,numeric
randomise(plate, seed, vector)

Arguments

plate

A Plate object.

seed

An integer number to set the seed (optional).

vector

A (usually random) vector to sort the plate with (optional).

Value

a Plate object in which the randomisation vector was stored in its metadata slot.

Note

Working with random seeds is tricky. In particular, pay attention that if you set the random seed just before running the randomise function, its results will become deterministic.

See also

Other Plate functions: Plate-class, PlateTypeToWellNames(), as.character,Well-method, nextWell(), plateMap(), plateWellVolume(), seekReagent(), setWell(), set_block(), sourceReagent()

Author

Charles Plessy

Examples

# Toy example of a 6-well plate containing a gradient increase of a reagent
p <- Plate(type = "6") |> set_block("A01~B03", "reagent", 1:6 * 10)
head(p)
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        10
#> A02        20
#> A03        30
#> B01        40
#> B02        50
#> B03        60

# Setting the same seed returns the same randomisation
randomise(p, seed = 1111) |> head()
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        40
#> A02        20
#> A03        50
#> B01        10
#> B02        60
#> B03        30
randomise(p, seed = 1111) |> head()
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        40
#> A02        20
#> A03        50
#> B01        10
#> B02        60
#> B03        30

# Not providing a seed returns a really random order
randomise(p) |> head()
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        20
#> A02        30
#> A03        50
#> B01        40
#> B02        60
#> B03        10
randomise(p) |> head()
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        30
#> A02        20
#> A03        40
#> B01        50
#> B02        10
#> B03        60

# The random order can also be passed with the vector argument
randomise(p, vector = sample(1:6)) |> head()
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        40
#> A02        10
#> A03        30
#> B01        60
#> B02        20
#> B03        50

# This means also that order can also be forced
randomise(p, vector = 6:1) |> head()
#> DataFrame with 6 rows and 1 column
#>       reagent
#>     <numeric>
#> A01        60
#> A02        50
#> A03        40
#> B01        30
#> B02        20
#> B03        10

# The randomisation vector is stored in the Plate object's metadata
p <- randomise(p)
p@metadata$randomisation_vector
#> [1] 3 6 2 1 4 5