choose_threshold | R Documentation |
This function makes for each patient the Penda test with different values for the treshold. It computes FDR, TPR and FPR for each patient and each threshold value.
choose_threshold(
controls,
L_H_list,
iterations,
simulation,
threshold_values,
quant_test = 0,
factor_test = 1
)
controls |
A matrix with genes expressions in controls for all the patients. |
L_H_list |
The list of lower and higher expressed genes matrices in the control. |
iterations |
The maximal number of iterations for the test. If the dysregulation list no longer changes, iterations are stop before. |
simulation |
The simulation with the list of initial data $initial_data and modified data in $simulated_data. |
threshold_values |
A vector with different values to test for the threshold. |
quant_test |
The quantile for the test. When L and H lists are void, we use the naive method. |
factor_test |
The factor for the test. The limit L will be quant(gene)/factor, and the limit H quant(gene)*factor. |
This function returns a matrix with 5 columns : the patient number, the value of the threshold tested, the FDR, the TPR and the FPR of the test.
# First, load and define the data, make the simulation
controls = penda::penda_data_ctrl[1:10, 1:13]
median_gene = apply(controls, 1, median, na.rm = TRUE)
median_gene = sort(median_gene)
controls = controls[names(median_gene), ]
simu_data = controls[,11:13]
controls = controls[,-(11:13)]
simulation = penda::simplified_simulation(simu_data, proportion = 0.3, threshold = 60)
# Second, lower and higher lists are computed
L_H_list = penda::compute_lower_and_higher_lists(controls, threshold = 0.99, s_max = 50)
# Third, choose the vector of threshold to test
threshold_values = c(0.1, 0.2, 0.3, 0.4, 0.5)
# Fourth, make the test on simulation for different thresholds
which_threshold = penda::choose_threshold(controls,
L_H_list,
iterations = 20,
simulation,
threshold_values,
quant_test = 0,
factor_test = 1)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.