Merge pull request #189 from MangiolaLaboratory/pass_fit_method_across_steps

track fit method in the attributes

Author: stemangiola

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: sccomp
 Title: Tests differences in cell-type proportion for single-cell data, robust to outliers
-Version: 1.99.15
+Version: 1.99.16
 Authors@R: c(person("Stefano", "Mangiola", email = "[email protected]",
                   role = c("aut", "cre"))
                   )

R/functions_multi_beta_binomial.R

@@ -329,6 +329,7 @@ sccomp_glm_data_frame_counts = function(.data,
     add_attr(formula_composition |> drop_environment(), "formula_composition") |>
     add_attr(formula_variability |> drop_environment(), "formula_variability") |>
     add_attr(parse_formula(formula_composition), "factors" ) |> 
+    add_attr(inference_method, "inference_method" ) |> 
 
     # Add class to the tbl
     add_class("sccomp_tbl") |> 

R/methods.R

@@ -609,7 +609,7 @@ sccomp_estimate.data.frame <- function(.data,
 sccomp_remove_outliers <- function(.estimate,
                                    percent_false_positive = 5,
                                    cores = detectCores(),
-                                   inference_method = "pathfinder",
+                                   inference_method = .estimate |> attr("inference_method"),
                                    output_directory = "sccomp_draws_files",
                                    verbose = TRUE,
                                    mcmc_seed = sample(1e5, 1),
@@ -639,7 +639,7 @@ sccomp_remove_outliers <- function(.estimate,
 sccomp_remove_outliers.sccomp_tbl = function(.estimate,
                                              percent_false_positive = 5,
                                              cores = detectCores(),
-                                             inference_method = "pathfinder",
+                                             inference_method = .estimate |> attr("inference_method"),
                                              output_directory = "sccomp_draws_files",
                                              verbose = TRUE,
                                              mcmc_seed = sample(1e5, 1),
@@ -959,6 +959,7 @@ sccomp_remove_outliers.sccomp_tbl = function(.estimate,
     add_attr(formula_composition, "formula_composition") |>
     add_attr(formula_variability, "formula_variability") |>
     add_attr(parse_formula(formula_composition), "factors" ) |> 
+    add_attr(inference_method, "inference_method" ) |> 
 
     # Add class to the tbl
     add_class("sccomp_tbl") |> 
@@ -1057,7 +1058,8 @@ sccomp_test.sccomp_tbl = function(.data,
   .count = .data |>  attr(".count")
   model_input = .data |> attr("model_input")
   truncation_df2 =  .data |>  attr("truncation_df2")
-
+  inference_method = .data |>  attr("inference_method")
+  
   # Abundance
   abundance_CI =
     get_abundance_contrast_draws(.data, contrasts)
@@ -1164,7 +1166,8 @@ sccomp_test.sccomp_tbl = function(.data,
 
     add_attr(.data |> attr("formula_composition"), "formula_composition") |>
     add_attr(.data |> attr("formula_variability"), "formula_variability") |>
-    
+    add_attr(inference_method, "inference_method" ) |> 
+
     # Add class to the tbl
     add_class("sccomp_tbl") 
 }

R/utilities.R

@@ -1820,7 +1820,7 @@ get_FDR = function(x){
 #' This function creates a series of 1D interval plots for cell-group effects, highlighting significant differences based on a given significance threshold.
 #'
 #' @param .data Data frame containing the main data.
-#' @param significance_threshold Numeric value specifying the significance threshold for highlighting differences. Default is 0.025.
+#' @param significance_threshold Numeric value specifying the significance threshold for highlighting differences. 
 #' @param test_composition_above_logit_fold_change A positive integer. It is the effect threshold used for the hypothesis test. A value of 0.2 correspond to a change in cell proportion of 10% for a cell type with baseline proportion of 50%. That is, a cell type goes from 45% to 50%. When the baseline proportion is closer to 0 or 1 this effect thrshold has consistent value in the logit uncontrained scale.
 #' @importFrom patchwork wrap_plots
 #' @importFrom forcats fct_reorder
@@ -1830,10 +1830,34 @@ get_FDR = function(x){
 #' 
 #' @return A combined plot of 1D interval plots.
 #' @examples
-#' # Example usage:
-#' # plot_1D_intervals(.data, "cell_group", 0.025, theme_minimal())
+#' 
+#' \donttest{
+#'   if (instantiate::stan_cmdstan_exists()) {
+#'     data("counts_obj")
+#'
+#'     estimate <- sccomp_estimate(
+#'       counts_obj,
+#'       ~ type,
+#'       ~1,
+#'       sample,
+#'       cell_group,
+#'       count,
+#'       cores = 1
+#'     ) |> 
+#'     sccomp_test()
+#'     
+#'   # Example usage:
+#'   my_plot = plot_1D_intervals(estimate)
+#'     
+#'   }
+#' }
+#'
+#' 
+
 plot_1D_intervals = function(.data, significance_threshold = 0.05, test_composition_above_logit_fold_change = .data |> attr("test_composition_above_logit_fold_change")){
 
+  message("sccomp says: Some FDR-significant populations may cross the fold change threshold. \n This because, as sccomp is a Bayesian method, the FDR is calculated according to Stephens (doi: 10.1093/biostatistics/kxw041), \n by sorting the probability of the null hypothesis in ascending order and calculating the cumulative average.")
+  
   # Define the variables as NULL to avoid CRAN NOTES
   parameter <- NULL
   estimate <- NULL

man/sccomp_remove_outliers.Rd

@@ -836,6 +836,7 @@ test_that("LOO", {
       .sample =  sample, 
       .cell_group = cell_group, 
       inference_method = "hmc",
+      cores = 1, 
       enable_loo = TRUE
     ) |> 
     expect_no_error()
@@ -848,6 +849,7 @@ test_that("LOO", {
       .sample =  sample, 
       .cell_group = cell_group, 
       inference_method = "hmc",
+      cores = 1, 
       enable_loo = TRUE
     ) |> 
     expect_no_error()
@@ -872,9 +874,10 @@ test_that("use two methods", {
       formula_composition = ~ 1, 
       .sample =  sample, 
       .cell_group = cell_group, 
-      inference_method = "hmc"
+      inference_method = "hmc",
+      cores = 1
     ) |> 
-    sccomp_remove_outliers(inference_method = "pathfinder") |> 
+    sccomp_remove_outliers(inference_method = "pathfinder", cores = 1) |> 
     expect_no_error()
 
 })