Aggregate data based on given grouping.

Description

aggregate aggregates data based on the specified aggregation method.

Usage

aggregate(
  population,
  variables,
  strata,
  operation = "mean",
  univariate = TRUE,
  ...
)

Arguments

Value

aggregated data of the same class as population.

Examples

population <- tibble::tibble(
  Metadata_group = c(
    "control", "control", "control", "control",
    "experiment", "experiment", "experiment",
    "experiment"
  ),
  Metadata_batch = c("a", "a", "b", "b", "a", "a", "b", "b"),
  AreaShape_Area = c(10, 12, 15, 16, 8, 8, 7, 7)
)
variables <- c("AreaShape_Area")
strata <- c("Metadata_group", "Metadata_batch")
aggregate(population, variables, strata, operation = "mean")

Remove redundant variables.

Description

correlation_threshold returns list of variables such that no two variables have a correlation greater than a specified threshold.

Usage

correlation_threshold(variables, sample, cutoff = 0.9, method = "pearson")

Arguments

Details

correlation_threshold is a wrapper for caret::findCorrelation.

Value

character vector specifying observation variables to be excluded.

Examples

suppressMessages(suppressWarnings(library(magrittr)))
sample <- tibble::tibble(
  x = rnorm(30),
  y = rnorm(30) / 1000
)

sample %<>% dplyr::mutate(z = x + rnorm(30) / 10)
variables <- c("x", "y", "z")

head(sample)
cor(sample)

# `x` and `z` are highly correlated; one of them will be removed

correlation_threshold(variables, sample)

Count the number of NAs per variable.

Description

count_na_rows counts the number of NAs per variable.

Usage

count_na_rows(population, variables)

Arguments

Value

data frame with frequency of NAs per variable.

Examples

population <- tibble::tibble(
  Metadata_group = c(
    "control", "control", "control", "control",
    "experiment", "experiment", "experiment", "experiment"
  ),
  Metadata_batch = c("a", "a", "b", "b", "a", "a", "b", "b"),
  AreaShape_Area = c(10, 12, 15, 16, 8, 8, 7, 7),
  AreaShape_length = c(2, 3, NA, NA, 4, 5, 1, 5)
)
variables <- c("AreaShape_Area", "AreaShape_length")
count_na_rows(population, variables)

Compute covariance matrix and vectorize.

Description

covariance computes the covariance matrix and vectorize it.

Usage

covariance(population, variables)

Arguments

Value

data frame of 1 row comprising vectorized covariance matrix.

Examples

population <- tibble::tibble(
  x = rnorm(30),
  y = rnorm(30),
  z = rnorm(30)
)

variables <- c("x", "y")

covariance(population, variables)

Remove variables with NA values.

Description

drop_na_columns returns list of variables which have greater than a specified threshold number of NAs.

Usage

drop_na_columns(population, variables, cutoff = 0.05)

Arguments

Value

character vector specifying observation variables to be excluded.

Examples

population <- tibble::tibble(
  Metadata_group = c(
    "control", "control", "control", "control",
    "experiment", "experiment", "experiment", "experiment"
  ),
  Metadata_batch = c("a", "a", "b", "b", "a", "a", "b", "b"),
  AreaShape_Area = c(10, 12, 15, 16, 8, 8, 7, 7),
  AreaShape_Length = c(2, 3, NA, NA, 4, 5, 1, 5)
)
variables <- c("AreaShape_Area", "AreaShape_Length")
drop_na_columns(population, variables)

Drop rows that are NA in all specified variables.

Description

drop_na_rows drops rows that are NA in all specified variables.

Usage

drop_na_rows(population, variables)

Arguments

Value

population without rows that have NA in all specified variables.

Examples

population <- tibble::tibble(
  Metadata_group = c(
    "control", "control", "control", "control",
    "experiment", "experiment", "experiment", "experiment"
  ),
  Metadata_batch = c("a", "a", "b", "b", "a", "a", "b", "b"),
  AreaShape_Area = c(10, 12, NA, 16, 8, 8, 7, 7),
  AreaShape_Length = c(2, 3, NA, NA, 4, 5, 1, 5)
)
variables <- c("AreaShape_Area", "AreaShape_Length")
drop_na_rows(population, variables)

Extract subpopulations.

Description

extract_subpopulations identifies clusters in the reference and population sets and reports the frequency of points in each cluster for the two sets.

Usage

extract_subpopulations(population, reference, variables, k)

Arguments

Value

list containing clusters centers (subpop_centers), two normalized histograms specifying frequency of each clusters in population and reference (subpop_profiles), and cluster prediction and distance to the predicted cluster for all input data (population_clusters and reference_clusters).

Examples

data <- tibble::tibble(
  Metadata_group = c(
    "control", "control", "control", "control",
    "experiment", "experiment", "experiment", "experiment"
  ),
  AreaShape_Area = c(10, 12, NA, 16, 8, 8, 7, 7),
  AreaShape_Length = c(2, 3, NA, NA, 4, 5, 1, 5)
)
variables <- c("AreaShape_Area", "AreaShape_Length")
population <- dplyr::filter(data, Metadata_group == "experiment")
reference <- dplyr::filter(data, Metadata_group == "control")
extract_subpopulations(
  population = population,
  reference = reference,
  variables = variables,
  k = 3
)

Generalized log transform data.

Description

generalized_log transforms specified observation variables using x = log( (x + sqrt(x ^ 2 + offset ^ 2 )) / 2 ).

Usage

generalized_log(population, variables, offset = 1)

Arguments

Value

transformed data of the same class as population.

Examples

population <- tibble::tibble(
  Metadata_Well = c("A01", "A02", "B01", "B02"),
  Intensity_DNA = c(8, 20, 12, 32)
)
variables <- c("Intensity_DNA")
generalized_log(population, variables)

A sparse matrix for sparse random projection.

Description

generate_component_matrix generates the sparse random component matrix for performing sparse random projection. If density is the density of the sparse matrix and n_components is the size of the projected space, the elements of the random matrix are drawn from

Usage

generate_component_matrix(n_features, n_components, density)

Arguments

Details

-sqrt(1 / (density * n_components)) with probability density / 2 0 with probability 1 - density sqrt(1 / (density * n_components)) with probability density / 2

Value

A sparse random matrix of size (n_features, n_components).

Examples

generate_component_matrix(500, 100, 0.3)

Normalize observation variables.

Description

normalize normalizes observation variables based on the specified normalization method.

Usage

normalize(
  population,
  variables,
  strata,
  sample,
  operation = "standardize",
  ...
)

Arguments

Value

normalized data of the same class as population.

Examples

suppressMessages(suppressWarnings(library(magrittr)))
population <- tibble::tibble(
  Metadata_group = c(
    "control", "control", "control", "control",
    "experiment", "experiment", "experiment", "experiment"
  ),
  Metadata_batch = c("a", "a", "b", "b", "a", "a", "b", "b"),
  AreaShape_Area = c(10, 12, 15, 16, 8, 8, 7, 7)
)
variables <- c("AreaShape_Area")
strata <- c("Metadata_batch")
sample <- population %>% dplyr::filter(Metadata_group == "control")
cytominer::normalize(population, variables, strata, sample, operation = "standardize")

Measure replicate correlation of variables.

Description

'replicate_correlation' measures replicate correlation of variables.

Usage

replicate_correlation(
  sample,
  variables,
  strata,
  replicates,
  replicate_by = NULL,
  split_by = NULL,
  cores = NULL
)

Arguments

Value

data frame of variable quality measurements

Examples

set.seed(123)
x1 <- rnorm(10)
x2 <- x1 + rnorm(10) / 100
y1 <- rnorm(10)
y2 <- y1 + rnorm(10) / 10
z1 <- rnorm(10)
z2 <- z1 + rnorm(10) / 1

batch <- rep(rep(1:2, each = 5), 2)

treatment <- rep(1:10, 2)

replicate_id <- rep(1:2, each = 10)

sample <-
  tibble::tibble(
    x = c(x1, x2), y = c(y1, y2), z = c(z1, z2),
    Metadata_treatment = treatment,
    Metadata_replicate_id = replicate_id,
    Metadata_batch = batch
  )

head(sample)

# `replicate_correlation`` returns the median, min, and max
# replicate correlation (across batches) per variable
replicate_correlation(
  sample = sample,
  variables = c("x", "y", "z"),
  strata = c("Metadata_treatment"),
  replicates = 2,
  split_by = "Metadata_batch",
  replicate_by = "Metadata_replicate_id",
  cores = 1
)

Reduce the dimensionality of a population using sparse random projection.

Description

sparse_random_projection reduces the dimensionality of a population by projecting the original data with a sparse random matrix. Generally more efficient and faster to compute than a Gaussian random projection matrix, while providing similar embedding quality.

Usage

sparse_random_projection(population, variables, n_components)

Arguments

Value

Dimensionality reduced population.

Examples

population <- tibble::tibble(
  Metadata_Well = c("A01", "A02", "B01", "B02"),
  AreaShape_Area_DNA = c(10, 12, 7, 7),
  AreaShape_Length_DNA = c(2, 3, 1, 5),
  Intensity_DNA = c(8, 20, 12, 32),
  Texture_DNA = c(5, 2, 43, 13)
)
variables <- c("AreaShape_Area_DNA", "AreaShape_Length_DNA", "Intensity_DNA", "Texture_DNA")
sparse_random_projection(population, variables, 2)

Feature importance based on data entropy.

Description

svd_entropy measures the contribution of each feature in decreasing the data entropy.

Usage

svd_entropy(variables, sample, cores = NULL)

Arguments

Value

data frame specifying the contribution of each feature in decreasing the data entropy. Higher values indicate more information.

Examples

sample <- tibble::tibble(
  AreaShape_MinorAxisLength = c(10, 12, 15, 16, 8, 8, 7, 7, 13, 18),
  AreaShape_MajorAxisLength = c(35, 18, 22, 16, 9, 20, 11, 15, 18, 42),
  AreaShape_Area = c(245, 151, 231, 179, 50, 112, 53, 73, 164, 529)
)
variables <- c("AreaShape_MinorAxisLength", "AreaShape_MajorAxisLength", "AreaShape_Area")
svd_entropy(variables, sample, cores = 1)

Transform observation variables.

Description

transform transforms observation variables based on the specified transformation method.

Usage

transform(population, variables, operation = "generalized_log", ...)

Arguments

Value

transformed data of the same class as population.

Examples

population <- tibble::tibble(
  Metadata_Well = c("A01", "A02", "B01", "B02"),
  Intensity_DNA = c(8, 20, 12, 32)
)
variables <- c("Intensity_DNA")
transform(population, variables, operation = "generalized_log")

Measure variable importance.

Description

variable_importance measures importance of variables based on specified methods.

Usage

variable_importance(
  sample,
  variables,
  operation = "replicate_correlation",
  ...
)

Arguments

Value

data frame containing variable importance measures.

Examples

set.seed(123)
x1 <- rnorm(10)
x2 <- x1 + rnorm(10) / 100
y1 <- rnorm(10)
y2 <- y1 + rnorm(10) / 10
z1 <- rnorm(10)
z2 <- z1 + rnorm(10) / 1

batch <- rep(rep(1:2, each = 5), 2)

treatment <- rep(1:10, 2)

replicate_id <- rep(1:2, each = 10)

sample <-
  tibble::tibble(
    x = c(x1, x2), y = c(y1, y2), z = c(z1, z2),
    Metadata_treatment = treatment,
    Metadata_replicate_id = replicate_id,
    Metadata_batch = batch
  )

head(sample)

# `replicate_correlation`` returns the median, min, and max
# replicate correlation (across batches) per variable
variable_importance(
  sample = sample,
  variables = c("x", "y", "z"),
  operation = "replicate_correlation",
  strata = c("Metadata_treatment"),
  replicates = 2,
  split_by = "Metadata_batch",
  replicate_by = "Metadata_replicate_id",
  cores = 1
)

Select observation variables.

Description

variable_select selects observation variables based on the specified variable selection method.

Usage

variable_select(
  population,
  variables,
  sample = NULL,
  operation = "variance_threshold",
  ...
)

Arguments

Value

variable-selected data of the same class as population.

Examples

# In this example, we use `correlation_threshold` as the operation for
# variable selection.

suppressMessages(suppressWarnings(library(magrittr)))
population <- tibble::tibble(
  x = rnorm(100),
  y = rnorm(100) / 1000
)

population %<>% dplyr::mutate(z = x + rnorm(100) / 10)

sample <- population %>% dplyr::slice(1:30)

variables <- c("x", "y", "z")

operation <- "correlation_threshold"

cor(sample)

# `x` and `z` are highly correlated; one of them will be removed

head(population)

futile.logger::flog.threshold(futile.logger::ERROR)

variable_select(population, variables, sample, operation) %>% head()

Remove variables with near-zero variance.

Description

variance_threshold returns list of variables that have near-zero variance.

Usage

variance_threshold(variables, sample)

Arguments

Details

variance_threshold is a reimplementation of caret::nearZeroVar, using the default values for freqCut and uniqueCut.

Value

character vector specifying observation variables to be excluded.

Examples

sample <- tibble::tibble(
  AreaShape_Area = c(10, 12, 15, 16, 8, 8, 7, 7, 13, 18),
  AreaShape_Euler = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0)
)
variables <- c("AreaShape_Area", "AreaShape_Euler")
variance_threshold(variables, sample)

Whiten data.

Description

whiten transforms specified observation variables by estimating a whitening transformation on a sample and applying it to the population.

Usage

whiten(population, variables, sample, regularization_param = 1)

Arguments

Value

transformed data of the same class as population.

Examples

population <- tibble::tibble(
  Metadata_Well = c("A01", "A02", "B01", "B02"),
  Intensity_DNA = c(8, 20, 12, 32),
  Texture_DNA = c(5, 2, 43, 13)
)
variables <- c("Intensity_DNA", "Texture_DNA")
whiten(population, variables, population, 0.01)