Fix moonlight z-score and updated datasets accordingly

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: Moonlight2R
 Type: Package
 Title: Identify oncogenes and tumor suppressor genes from omics data
-Version: 1.11.0
+Version: 1.11.1
 Date:
 Authors@R:
   c(person("Mona", "Nourbakhsh",
@@ -90,7 +90,7 @@ SystemRequirements: CScapeSomatic
 VignetteBuilder: knitr
 URL: https://github.com/ELELAB/Moonlight2R
 BugReports: https://github.com/ELELAB/Moonlight2R/issues
-RoxygenNote: 7.3.3
 LazyData: false
 Encoding: UTF-8
 Config/testthat/edition: 3
+Config/roxygen2/version: 8.0.0

NEWS.md

@@ -1,3 +1,9 @@
+# Moonlight2R 1.11.1
+
+## Summary
+
+* Fix FEA Moonlight Z-score to consider experimental logFC in the score calculation
+
 # Moonlight2R 1.9.4
 
 ## Summary

R/FEA.R

@@ -25,7 +25,7 @@ FEA <- function(BPname = NULL,
 
   # List of variable names
   variables_to_check <- c("DiseaseList", "EAGenes")
-
+  
   # Check and load variables if they do not exist
   for (variable_name in variables_to_check) {
     if (! variable_name %in% names(.GlobalEnv)) {
@@ -74,7 +74,6 @@ FEA <- function(BPname = NULL,
     # sorting strategy to avoid different sorting in case of tie ranks
     rankings <- rankings[order(-rankings, names(rankings))]
 
-
     pathwayNamesList <- list()
     for (k in seq_along(lf2)) {
 
@@ -101,11 +100,24 @@ FEA <- function(BPname = NULL,
       res <- as.data.frame(matrix(0, nrow = 1, ncol = 2,
                                   dimnames = list(1, c("pathway",
                                                      "Moonlight.Z.score"))))
+
+
+
+      pathway_data <- fgseaRes[fgseaRes$pathway == lf2[k],]
+
+      leadingEdge_genes <- unlist(pathway_data$leadingEdge)
 
       selected_diseases <- as.data.frame(DiseaseList[[which(names(DiseaseList) == lf2[k])]])
 
+      selected_diseases <- selected_diseases[selected_diseases$ID %in% leadingEdge_genes, ]
+
       res$pathway <- lf2[k]
 
+      selected_diseases$logFC <- DiffMatrix$logFC[match(selected_diseases$ID, rownames(DiffMatrix))]
+
+      selected_diseases <- selected_diseases[!is.na(selected_diseases$logFC), ]
+
+
       Zscore <- .compute_moonlight_zscore(selected_diseases)
 
       res$Moonlight.Z.score <- Zscore
@@ -124,7 +136,7 @@ FEA <- function(BPname = NULL,
       rename("Diseases.or.Functions.Annotation" = pathway, "p.value" = pval) %>% 
       select("Diseases.or.Functions.Annotation", "Moonlight.Z.score","p.value", "padj", "ES", "NES", "size", "leadingEdge")
 
-
+    
   }
   else if (method == "ora") {
 
@@ -173,37 +185,17 @@ FEA <- function(BPname = NULL,
       }
 
       GeneList <- GeneList[GeneList$PROBE_ID %in% selected_diseases$ID, ]
-
+      
       selected_diseases <- selected_diseases[selected_diseases$ID %in% GeneList[, "PROBE_ID"], ]
-      selected_diseases[, "Exp.Log.Ratio"] <- gsub(",", ".", selected_diseases[, "Exp.Log.Ratio"])
-      selected_diseases[, "Exp.Log.Ratio"] <- as.numeric(selected_diseases[, "Exp.Log.Ratio"])
 
-      rownames(selected_diseases) <- selected_diseases$ID
+      selected_diseases$logFC <- GeneList$logFC[match(selected_diseases$ID, GeneList$PROBE_ID)]
+
+      selected_diseases <- selected_diseases[!is.na(selected_diseases$logFC), ]
 
+      rownames(selected_diseases) <- selected_diseases$ID
+
       res$Molecules <- paste0(GeneList$PROBE_ID, collapse = ",")
 
-      for (idx in seq.int(nrow(selected_diseases))) {
-
-        currTR <- selected_diseases$ID[idx]
-
-        if (length(grep("Increases", selected_diseases[currTR, "Findings"])) == 1) {
-
-          if (sign(GeneList[currTR, "logFC"]) > 0) {
-            selected_diseases[currTR, "Prediction..based.on.expression.direction."] <- "Increased"
-          } else if (sign(GeneList[currTR, "logFC"]) < 0) {
-            selected_diseases[currTR, "Prediction..based.on.expression.direction."] <- "Decreased"
-          }
-        }
-
-        if (length(grep("Decreases", selected_diseases[currTR, "Findings"])) == 1) {
-          if (sign(GeneList[currTR, "logFC"]) < 0) {
-            selected_diseases[currTR, "Prediction..based.on.expression.direction."] <- "Increased"
-          } else if (sign(GeneList[currTR, "logFC"]) > 0) {
-            selected_diseases[currTR, "Prediction..based.on.expression.direction."] <- "Decreased"
-          }
-        }
-      }
-
       Zscore <- .compute_moonlight_zscore(selected_diseases)
 
       res$Moonlight.Z.score <- Zscore
@@ -237,22 +229,25 @@ FEA <- function(BPname = NULL,
 #' @noRd
 .compute_moonlight_zscore <- function(selected_diseases) {
 
+  if(!all(c("Findings", "logFC") %in% colnames(selected_diseases))){
+    stop("The input DiseaseList does not contain 'Findings' and 'logFC' columns for Moonlight z-score calculations")
+  }
+
   Correlation <- matrix(0, nrow(selected_diseases), 1)
 
   selected_diseases <- cbind(selected_diseases, Correlation)
-
-  if (length(grep("Decreases", selected_diseases$Findings)) != 0) {
+  if (any(grepl("Increases|Decreases", selected_diseases$Findings, ignore.case = TRUE))){
 
         selected_diseases[grep("Decreases", selected_diseases$Findings), "Findings"] <- -1
         selected_diseases[grep("Increases", selected_diseases$Findings), "Findings"] <- 1
         selected_diseases[grep("Affects", selected_diseases$Findings), "Findings"] <- 0
         selected_diseases[, "Findings"] <- as.numeric(selected_diseases[, "Findings"])
 
-        selected_diseases[, "Exp.Log.Ratio"] <- gsub(",", ".", selected_diseases[, "Exp.Log.Ratio"])
-        selected_diseases[, "Exp.Log.Ratio"] <- as.numeric(selected_diseases[, "Exp.Log.Ratio"])
-        
-        PredictionIncreased <- which(sign(selected_diseases$Exp.Log.Ratio) == selected_diseases$Findings)
-        PredictionDecreased <- which(sign(selected_diseases$Exp.Log.Ratio) != selected_diseases$Findings)
+        selected_diseases[, "logFC"] <- gsub(",", ".", selected_diseases[, "logFC"])
+        selected_diseases[, "logFC"] <- as.numeric(selected_diseases[, "logFC"])
+
+        PredictionIncreased <- which(sign(selected_diseases$logFC) == selected_diseases$Findings)
+        PredictionDecreased <- which(sign(selected_diseases$logFC) != selected_diseases$Findings)
         PredictionAffected <- which(sign(selected_diseases$Findings) == 0)
 
         selected_diseases[PredictionIncreased, "Correlation"] <- 1
@@ -263,7 +258,7 @@ FEA <- function(BPname = NULL,
       } else {
         Zscore <- 0
       }
-
+      
       return(Zscore)
 
 }

R/PRA.R

@@ -13,7 +13,7 @@
 #' data(DiseaseList)
 #' data(tabGrowBlock)
 #' data(knownDriverGenes)
-#' dataPRA <- PRA(dataURA = dataURA[seq.int(2),],
+#' dataPRA <- PRA(dataURA = dataURA,
 #' BPname = c("apoptosis","proliferation of cells"),
 #' thres.role = 0)
 PRA <- function(dataURA,

tests/testthat/test-TFinfluence.R

@@ -9,17 +9,17 @@ data(dataMAVISp)
 
 
 reference_TFresults <- data.frame(
-    Target = c("CHST1", "GEN1", "TAP1", "TAP1"),
-    Moonlight_gene_z_score = c(1.06066017177982, 1.30096115353815, 0.8262079533974, 0.8262079533974),
-    Moonlight_Oncogenic_Mediator = c("TSG", "OCG", "OCG", "OCG"),
-    logFC_target = c(1.94917640749552, 1.36435629696216, 1.36486453774995, 1.36486453774995),
-    TF = c(NA, NA, "IRF1", "IRF2"),
-    InteractionType = c(NA, NA, "Activation", "Activation"),
-    PMID = c(NA, NA, "18694960", "15778351"),
+    Target = c("PDE2A", "GCAT", "ARMC9", "FAM155A"),
+    Moonlight_gene_z_score = c(1.37642904406401, 0.810977174901446, 0.367466020609254, 1.19863281130568),
+    Moonlight_Oncogenic_Mediator = c("TSG", "TSG", "TSG", "OCG"),
+    logFC_target = c(-3.36278468754307, 1.0407914607103, 1.03363210942098, 1.97338960700756),
+    TF = as.character(c(NA, NA, NA, NA)),
+    InteractionType = as.character(c(NA, NA, NA, NA)),
+    PMID = as.character(c(NA, NA, NA, NA)),
     tf_mutation = as.character(c(NA, NA, NA, NA)),
     stab_class = as.character(c(NA, NA, NA, NA)),
     in_MAVISp = c(FALSE,FALSE,FALSE,FALSE),
-    mutation_available = c(FALSE,FALSE,FALSE,FALSE),
+    mutation_available = c(FALSE, FALSE, FALSE, FALSE),
     stringsAsFactors = FALSE
 )