Log2 pre processing (#315)

* add Log2 as preprocessing
* Added lines to help files and report output

docs/interface-details/pre-processing.md

@@ -13,10 +13,10 @@ The Pre-processing tab is divided into two main sections: the side panel and the
 
 In the side panel, you have the following options:
 
--   **Pre-Processing Procedures**: You can select from various pre-processing procedures.
 
-    -   Options: none, filterOnly, vst_DESeq, simpleCenterScaling, Scaling_0_1, log10, pareto_scaling, ln
-    -   vst_DESeq also requires the selection of a design formula.
+- **Pre-Processing Procedures**: You can select from various pre-processing procedures.
+  - Options: none, filterOnly, vst_DESeq, simpleCenterScaling, Scaling_0_1, log10, log2, pareto_scaling, ln
+  - vst_DESeq also requires the selection of a design formula.
 
 -   **Select Batch Effect Column**: Choose a batch effect if applicable. Possible choices are taken from the sample annotation columns.
 

program/shinyApp/R/Report_text_snippets.R

@@ -80,6 +80,8 @@ snippet_preprocessing <- function(
     snippet <- paste0(snippet, "The base-10 logarithm of each data point was calculated. If a single zero value was present, 1 was added to all points to avoid undefined results.\n")
   } else if (params$PreProcessing_Procedure == "ln") {
     snippet <- paste0(snippet, "The natural logarithm of each data point was calculated. If a single zero value was present, 1 was added to all points to avoid undefined results.\n")
+  } else if (params$PreProcessing_Procedure == "log2") {
+    snippet <- paste0(snippet, "The base-2 logarithm of each data point was calculated. If a single zero value was present, 1 was added to all points to avoid undefined results.\n")
   } else if (params$PreProcessing_Procedure == "pareto_scaling") {
     snippet <- paste0(snippet, "The data was parteo scaled. Pareto scaling emphasizes the importance of small values by dividing each data point by the square root of its standard deviation.\n")
   } else if(params$PreProcessing_Procedure == "None") {

program/shinyApp/R/pre_processing/ui.R

@@ -14,6 +14,7 @@ pre_processing_sidebar_panel <- sidebarPanel(
       "centering to 0 and scaling" = "simpleCenterScaling",
       "scaling values to be within 0 and 1" = "Scaling_0_1",
       "log10" = "log10",
+      "log2" = "log2",
       "Pareto scaling (mean-centered and scaled by the square root of the standard deviation)" = "pareto_scaling",
       "natural logarithm" = "ln"
     ),

program/shinyApp/R/pre_processing/util.R

@@ -10,7 +10,7 @@ preprocessing <- function(data, omic_type, procedure){
   if(procedure %in% c("Scaling_0_1", "pareto_scaling")){
     return(scaling_normalisation(data, omic_type, procedure))
   }
-  if(procedure %in% c("log10", "ln")){
+  if(procedure %in% c("log10", "ln", "log2")){
     return(ln_normalisation(data, omic_type, procedure))
   }
   if(procedure == "none"){
@@ -72,7 +72,18 @@ scaling_normalisation <- function(data, omic_type, scaling_procedure){
 
 ln_normalisation <- function(data, omic_type, logarithm_procedure){
   # Center and scale the data
-  logarithm <- ifelse(logarithm_procedure == "log10", log10, log)
+  if(logarithm_procedure == "log10")
+  {
+    logarithm = log10
+  }
+  else if(logarithm_procedure == "log2")
+  {
+    logarithm = log2
+  }
+  else
+  {
+    logarithm = log
+  }
   # prefilter the data
   data <- prefiltering(data, omic_type)
   # log the data and always add 1 to avoid -Inf

program/shinyApp/helpfiles/PreProcessing_Procedures.md

@@ -55,7 +55,13 @@
   - Special consideration is given to handling zero values to avoid undefined 
     results: If any zero values are present, +1 is added to all values before applying
     the logarithm.
-
+
+- **Logarithm Base 2 (log2):**
+  - The base-2 logarithm of each data point is calculated.
+  - Special consideration is given to handling zero values to avoid undefined 
+    results: If any zero values are present, +1 is added to all values before applying
+    the logarithm.
+
 - **Pareto Scaling:**
   - Pareto scaling emphasizes the importance of small values by dividing each data point by the square root of its standard deviation.
   - This method is suitable for datasets with a wide range of values.