v0.1.5

Author: bonagurol

.Rhistory

@@ -2,4 +2,6 @@
 condor.Rproj
 .Rproj.user/
 inst/doc
-.test_files
+.test_files
+docs
+.Rhistory

.gitignore

@@ -1,14 +1,16 @@
-Package: condor
+Package: cyCONDOR
 Type: Package
 Title: Flow Cytometry data analysis toolbox
-Version: 0.1.4
+Version: 0.1.5
 Author: Lorenzo Bonaguro
 Maintainer: Lorenzo Bonaguro <lorenzobonaguro@uni-bonn.de>
 Description: Flow cytometry analysis workflow. The aim of this project if to 
     provide an intuitive workflow for the analysis of high-dimensionality cytometry data in R.
+URL: https://github.com/lorenzobonaguro/condor
 License: GPL-3 + file LICENSE
 Encoding: UTF-8
 LazyData: FALSE
+biocViews:
 Imports: 
     ggplot2, 
     RColorBrewer,
@@ -45,7 +47,8 @@ Imports:
     Hmisc,
     DelayedMatrixStats,
     cowplot,
-    dplyr
+    dplyr,
+    reticulate
 RoxygenNote: 7.2.3
 VignetteBuilder: knitr
 Suggests: 

CHANGELOG

@@ -7,9 +7,12 @@ export(barplot_frequency)
 export(boxplot_and_stats)
 export(change_param_name)
 export(check_IDs)
+export(clr)
 export(confusionMatrix)
 export(confusion_HM)
+export(densityplot_marker)
 export(df_frequency)
+export(dotplot_cyto)
 export(filter_fcd)
 export(harmonize_PCA)
 export(harmonize_intensities)
@@ -24,6 +27,8 @@ export(predict_labels)
 export(prepFcsFolderData)
 export(prep_fcd)
 export(prep_fjw)
+export(runAstir_cellstates)
+export(runAstir_celltype)
 export(runDM)
 export(runFlowSOM)
 export(runPCA)
@@ -65,9 +70,11 @@ import(randomForest)
 import(readr)
 import(readxl)
 import(reshape2)
+import(reticulate)
 import(slingshot)
 import(stringr)
 import(umap)
 importFrom(Matrix,sparseMatrix)
 importFrom(igraph,membership)
 importFrom(utils,packageDescription)
+importFrom(utils,write.csv)

DESCRIPTION

@@ -0,0 +1,75 @@
+# cyCONDOR 0.1.5
+
+* Fix bug in the definition of tab separator when loading csv files
+* Added clr transformation for CITE-seq data together with minor improvements to the transformation function
+* Add clustering option in confusion matrix
+* Added visualization of 2D plots of PCA
+* Option to export plots as raster
+* Added function to plot conventional flow 2d plots
+* Added function to plot a density plot for marker expression
+* Implementation of Astir (with Python)
+* Restructured documentation and vignette
+* Added GitPages website with documentation and tutorials
+* Name change to `cyCONDOR`
+* Edited the pseudotime function to run in a loop for the starting clusters
+
+# cyCONDOR 0.1.4
+
+## Reference name: condor 0.1.4
+
+* Add ML classifier with CytoML
+* Included Hmisc as requirment for Violin plto marker function
+* Fix bug with UMAP plotting function when faceting (default for facet_by_fariable set to FALSE not NULL)
+* Fixed package loading message
+* Tested diffusion map and imporoved function
+* Tested pseutodime and improved function
+* Splittied functions in multiple files to make them easier to find
+* Improved package documentation
+* Improved package vignette
+
+# cyCONDOR 0.1.3
+
+## Reference name: condor 0.1.3
+
+* FlowSOM function can retain the model to plot the SOM tree afterwards
+* Added Function to read flowjo workspaces to a condor campatible format
+* Fixed filter function when the 'extra' slot is occupied
+
+# cyCONDOR 0.1.2
+
+## Reference name: condor 0.1.2
+
+* Several bug fixes
+* Added function to calculate tSNE dimensionality reduction
+* Possibility to limit the number of PC used for clustering and non-linear dimensionality reduction
+* Added function to calculate Pseudotime (slignshot)
+* Included an easy-to-export differential frequency table
+* Added a function for random subsetting of the dataset
+* Added visualization of PC loadings
+* Added Pseudobulk PCA Analysis
+* Added function to easily export cellular frequency
+* Added function to change the parameter names (of the fcs files)
+* Added function to visualize DRs as density plot
+* Added Violin plot visualization of marker expression
+* Included option to not cluster rows and columns in heatmaps
+* Included option to show cluster numbers in the dotplot
+* Add multicore support to tSNE
+* Added workflow for UMAP projection and label transfer
+
+# cyCONDOR 0.1.1 
+
+## Reference name: condor 0.1.1
+
+* Add function to merge condor objects
+* Updated LoadFCS function to be fully compatible with .csv files
+* Updated UMAP function to run on multiple cores
+* Updated RPhenograph function to run on multiple cores
+* Added function to run FlowSOM clustering
+* Fix issues with UMAP parametes selections
+* Added option to specifiy the delimiter for the csv files
+
+# cyCONDOR 0.1.0
+
+## Reference name: condor 0.1.0
+
+* Initial release

NAMESPACE

@@ -0,0 +1,139 @@
+#' runAstir_celltype
+#'
+#' @title runAstir_celltype
+#' @description Predict cell types using Astir. This package requires the python library `astir` and `reticulate` to work.
+#' @param fcd Flow cytometry dataset.
+#' @param data_slot Data slot to use for the analysis (e.g. "orig" or "norm").
+#' @param analysis_path Full path to the output folder of astir analysis.
+#' @param manifest_name Filename of the manifest file, this file must be located in the `analysis_path` folder.
+#' @param max_epochs Maximum number of epochs.
+#' @param learning_rate Learning Rate.
+#' @param initial_epochs Initial Epochs.
+#' @import reticulate
+#' @return runAstir_celltype
+#'
+#' @export
+runAstir_celltype <- function(fcd,
+                              data_slot,
+                              analysis_path,
+                              manifest_name,
+                              max_epochs,
+                              learning_rate,
+                              initial_epochs) {
+
+  # Save the expression matrix as csv
+  write.csv(x = fcd$expr[[data_slot]], file = paste0(analysis_path, "expr.csv"))
+
+  # Import the astir python package
+  ast_fun <- reticulate::import("astir")
+
+  # Set the directories for the analysis
+  expr_path <- paste0(analysis_path, "expr.csv")
+
+  manifest_path <- paste0(analysis_path, manifest_name)
+
+  ast <- ast_fun$from_csv_yaml(csv_input = expr_path, marker_yaml = manifest_path)
+
+  batch_size = dim(ast$get_type_dataset()$get_exprs_df())[1]/100
+
+  ast$fit_type(max_epochs = as.integer(max_epochs),
+               batch_size = as.integer(batch_size),
+               learning_rate = learning_rate,
+               n_init_epochs = as.integer(initial_epochs))
+
+  print(table(ast$get_celltypes()))
+
+  probabilities <- ast$get_celltype_probabilities()
+
+  diagnostic <- ast$diagnostics_celltype()
+
+  cell_types <- ast$get_celltypes()
+
+  write.csv(x = probabilities, file = paste0(analysis_path, "probabilities.csv"))
+
+  write.csv(x = diagnostic, file = paste0(analysis_path, "diagnostic.csv"))
+
+  write.csv(x = cell_types, file = paste0(analysis_path, "cell_types.csv"))
+
+  # Add the cell type prediction in the condor object
+
+  # Prepare the dataframe
+  df <- data.frame(cell_type = cell_types,
+                   Description = paste0("Astir_cell_type_", data_slot,
+                                        "_Max_Epoc_", max_epochs,
+                                        "Learning_Rate_", learning_rate,
+                                        "_Initial_Epochs_", initial_epochs))
+
+  fcd[["astir"]][[paste0("Astir_cell_type_", data_slot)]] <- df
+
+  return(fcd)
+
+}
+
+#' runAstir_cellstates
+#'
+#' @title runAstir_cellstates
+#' @description Predict cell states using Astir. This package requires the python library `astir` and `reticulate` to work.
+#' @param fcd Flow cytometry dataset.
+#' @param data_slot Data slot to use for the analysis (e.g. "orig" or "norm").
+#' @param analysis_path Full path to the output folder of astir analysis.
+#' @param manifest_name Filename of the manifest file, this file must be located in the `analysis_path` folder.
+#' @param max_epochs Maximum number of epochs.
+#' @param learning_rate Learning Rate.
+#' @param initial_epochs Initial Epochs.
+#' @import reticulate
+#' @importFrom utils write.csv
+#' @return runAstir_cellstates
+#'
+#' @export
+runAstir_cellstates <- function(fcd,
+                                data_slot,
+                                analysis_path,
+                                manifest_name,
+                                max_epochs,
+                                learning_rate,
+                                initial_epochs) {
+
+  # Save the expression matrix as csv
+  write.csv(x = fcd$expr[[data_slot]], file = paste0(analysis_path, "expr.csv"))
+
+  # Import the astir python package
+  ast_fun <- reticulate::import("astir")
+
+  # Set the directories for the analysis
+  expr_path <- paste0(analysis_path, "expr.csv")
+
+  manifest_path <- paste0(analysis_path, manifest_name)
+
+  # Parpare to run astir
+  ast <- ast_fun$from_csv_yaml(csv_input = expr_path, marker_yaml = manifest_path)
+
+  batch_size = dim(ast$get_type_dataset()$get_exprs_df())[1]/100
+
+  ast$fit_state(max_epochs = as.integer(max_epochs),
+                batch_size = as.integer(batch_size),
+                learning_rate = learning_rate,
+                n_init_epochs = as.integer(initial_epochs))
+
+  diagnostic <- ast$diagnostics_cellstate()
+
+  cell_states <- ast$get_cellstates()
+
+  write.csv(x = diagnostic, file = paste0(analysis_path, "diagnostic.csv"))
+
+  write.csv(x = cell_states, file = paste0(analysis_path, "cell_states.csv"))
+
+  # Add the cell type prediction in the condor object
+
+  # Prepare the dataframe
+  df <- data.frame(cell_states,
+                   Description = paste0("Astir_cell_state_", data_slot,
+                                        "_Max_Epoc_", max_epochs,
+                                        "Learning_Rate_", learning_rate,
+                                        "_Initial_Epochs_", initial_epochs))
+
+  fcd[["astir"]][[paste0("Astir_cell_state_", data_slot)]] <- df
+
+  return(fcd)
+
+}