tests/testthat/test_z_fit_equations.R

In Guilz/rfret: Analyze FRET Binding Data

context("Curve fitting and equations")

# Test FRET pipeline
# Load pre-computed expected results
load(file = "./fret/fret_expected_fits.rda")

# Run example dataset through the entire pipeline and keep resulting fit objects
fret_data_for_fit <- fret_binding_data %>%
    format_data(data_type = "fret") %>%
    fret_average_replicates() %>%
    fret_correct_signal()

fret_hyperbolic_fit <- fret_data_for_fit %>%
    fit_binding_model(binding_model = "hyperbolic")

fret_hill_fit <- fret_data_for_fit %>%
    fit_binding_model(binding_model = "hill")

fret_quadratic_fit <- fret_data_for_fit %>%
    fit_binding_model(binding_model = "quadratic", probe_concentration = 5)

# Test that results match expected precomputed results
test_that("fit_binding_model gives correct results on fret_binding_data", {
    expect_equal(object = fret_hyperbolic_fit,
                 expected = fret_expected_hyperbolic_fit)
    expect_equal(object = fret_hill_fit,
                 expected = fret_expected_hill_fit)
    expect_equal(object = fret_quadratic_fit,
                 expected = fret_expected_quadratic_fit)
})

# Test FP/FA pipeline
# Load pre-computed expected results
load(file = "./fp/fp_expected_fits.rda")

# Run example dataset through the entire pipeline and keep resulting fit objects
# Using polarization signal
fp_data_for_fit <- fp_binding_data %>%
    format_data(data_type = "fp") %>%
    fp_average_replicates() %>%
    fp_use_signal("polarization")

fp_hyperbolic_fit <- fp_data_for_fit %>%
    fit_binding_model(binding_model = "hyperbolic")

fp_hill_fit <- fp_data_for_fit %>%
    fit_binding_model(binding_model = "hill")

fp_quadratic_fit <- fp_data_for_fit %>%
    fit_binding_model(binding_model = "quadratic", probe_concentration = 3)

# Using anisotropy signal
fa_data_for_fit <- fp_binding_data %>%
    format_data(data_type = "fp") %>%
    fp_average_replicates() %>%
    fp_use_signal("anisotropy")

fa_hyperbolic_fit <- fa_data_for_fit %>%
    fit_binding_model(binding_model = "hyperbolic")

fa_hill_fit <- fa_data_for_fit %>%
    fit_binding_model(binding_model = "hill")

fa_quadratic_fit <- fa_data_for_fit %>%
    fit_binding_model(binding_model = "quadratic", probe_concentration = 3)

# Test that results match expected precomputed results
test_that("fit_binding_model gives correct results on fp_binding_data", {
    expect_equal(object = fp_hyperbolic_fit,
                 expected = fp_expected_hyperbolic_fit)
    expect_equal(object = fp_hill_fit,
                 expected = fp_expected_hill_fit)
    expect_equal(object = fp_quadratic_fit,
                 expected = fp_expected_quadratic_fit)
    expect_equal(object = fa_hyperbolic_fit,
                 expected = fa_expected_hyperbolic_fit)
    expect_equal(object = fa_hill_fit,
                 expected = fa_expected_hill_fit)
    expect_equal(object = fa_quadratic_fit,
                 expected = fa_expected_quadratic_fit)
})

# Test that an unknown binding model gives an error.
my_failing_command <- rlang::quo(
    fret_data_for_fit %>%
        fit_binding_model(binding_model = "unknown")
)

test_that("fit_binding_model detects unknown models", {
    expect_error(object = rlang::eval_tidy(my_failing_command))
})