Add files via upload

Author: tanlabcode

R/BenchMark/ESC_human/analysis_library.R

@@ -1,5 +1,11 @@
+# calculate the variance across the rows or the columns
 MatVar <- function(x, dim = 1, ...) {
 
+  # parameter in the function
+  # x: the input matrix 
+  # dim: determine the variance across the row or column.
+  # 1 is the one across the row and 2 is the one across the column
+  
   if(dim == 1){
 
     rowSums((x - rowMeans(x, ...))^2, ...)/(dim(x)[2] - 1)
@@ -11,22 +17,34 @@ MatVar <- function(x, dim = 1, ...) {
   } else stop("Please enter valid dimension")
 }
 
+# plot the pathway ratio
 plot_pathway_ratio <- function(dataV3, dataType, pathway_name){
 
+  # Parameter in the function
+  # dataV3: the ratio of the different methods, the columns are 
+  # the different methods
+  # dataType: the cell types
+  # pathway_name: the name of the pathway database
+  
   if(dataType == "NPC"){
-    # select the raw data, the imputed data by scImpute, scbMC, MAGIC
+    
+    # define the number of data points
     N = dim(dataV3)[2]
 
+    # define the lowest levels of the pvalues
     hlim = 190
 
+    # prepare the data as a vector for the boxplot
     dataV1 = data.frame(y = (as.vector(as.matrix(dataV3))))
 
     dataV1$group = rep(c(1:5),N)
 
+    # calculate the pvalues
     my_comparisons = list( c("1", "5"), c("2", "5"), c("3", "5"), c("4", "5"))
 
     pval = compare_means(y ~ group,data = dataV1, method = "t.test", ref.group = "5")
 
+    # plot the boxplots
     pl = ggboxplot(dataV1, x = "group", y = "y", fill = "group",
                     palette = c("#00AFBB","#0000CD", "#E7B800", "#FC4E07", "#6ebb00"),outlier.shape = NA) +
       stat_boxplot(geom = "errorbar", width = 0.3) + 
@@ -45,19 +63,23 @@ plot_pathway_ratio <- function(dataV3, dataType, pathway_name){
 
   }else{
 
-    # select the raw data, the imputed data by scImpute, scbMC, MAGIC
+    # define the number of data points
     N = dim(dataV3)[2]
 
+    # define the lowest levels of the pvalues
     hlim = 150
 
+    # prepare the data as a vector for the boxplot
     dataV1 = data.frame(y = (as.vector(as.matrix(dataV3))))
 
+    # calculate the pvalues
     dataV1$group = rep(c(1:5),N)
 
     my_comparisons = list( c("1", "5"), c("2", "5"), c("3", "5"), c("4", "5"))
 
     pval = compare_means(y ~ group,data = dataV1, method = "t.test", ref.group = "5")
 
+    # plot the boxplots
     pl = ggboxplot(dataV1, x = "group", y = "y", fill = "group",
                     palette = c("#00AFBB","#0000CD", "#E7B800", "#FC4E07", "#6ebb00"),outlier.shape = NA) +
       stat_boxplot(geom = "errorbar", width = 0.3) + 
@@ -83,38 +105,48 @@ plot_pathway_ratio <- function(dataV3, dataType, pathway_name){
 
 generate_index <- function(data_set, pathway_name){
 
+  # Parameter in the function
+  # data_set: the data of cell types
+  # pathway_name: the name of pathway database
+  
   # load the gene
-  data_gene = fread(file = "data_all/gene_ESC.csv", 
-                     header = FALSE)
+  data_gene = fread(file = "data_all/gene_name.csv", 
+                    sep=',', header = FALSE)
 
+  # load the scRNAseq data
   data_sc = as.matrix(fread(file = paste0("data_all/data_sc_",
                                            data_set,
                                            ".csv")))
 
-  
+  # calculate the variance across the genes
   var0 = MatVar(data_sc,1)
 
   index_sc = var0 > 1e-10
 
   data_gene = data_gene[index_sc,]
 
-  
   n_gene = dim(data_gene)[1]
 
-  
+  # determine the pathway numbers
   if(pathway_name == "IPA"){
+    
     N = 186
+    
   }
 
   if(pathway_name == "KEGG"){
+    
     N = 186
+    
   }
 
   if(pathway_name == "REACTOME"){
+    
     N = 674
+    
   }
 
-  
+  # calculate the index
   index = list()
 
   for(i in c(1:N)){
@@ -154,9 +186,13 @@ generate_index <- function(data_set, pathway_name){
   )
 } 
 
-
+# Calculate the ratio
 calculate_ratio <- function(data_set, pathway_name, method_name){
 
+  # Parameter in the function
+  # data_set: the data of cell type
+  # pathway_name: the name of the pathway
+  # method_name: the name of the method used for the imputation
 
   load(file = paste0("data_all/",
                      pathway_name,
@@ -203,19 +239,23 @@ calculate_ratio <- function(data_set, pathway_name, method_name){
                                      "_scrabble_imputation.rds"))
   }
 
+  # determine the pathway numbers
   if(pathway_name == "IPA"){
 
     N = 186
+    
   }
 
   if(pathway_name == "KEGG"){
 
     N = 186
+    
   }
 
   if(pathway_name == "REACTOME"){
 
     N = 674
+    
   }
 
   values = c()
@@ -231,6 +271,7 @@ calculate_ratio <- function(data_set, pathway_name, method_name){
     mean_value = c()
 
     if (N_index > 10){
+      
       for(j in c(1:101)){
 
         tmp_data = data_sc[tmp_index[j,],]