add fraction matching based on only the light / na labels

Author: tonywu1999

R/utils_fractions.R

@@ -265,11 +265,11 @@
 #' @keywords internal
 .removeOverlappingFeatures = function(input) {
     fraction_keep = Fraction = NULL
-    
+
     if (data.table::uniqueN(input$Fraction) > 1) {
-        input[, fraction_keep := .getCorrectFraction(.SD), 
-              by = "feature", 
-              .SDcols = c("feature", "Fraction", "Run", "Intensity")]
+        input[, fraction_keep := .getCorrectFraction(.SD),
+              by = "feature",
+              .SDcols = c("feature", "Fraction", "Run", "Intensity", "IsotopeLabelType")]
         input = input[Fraction == fraction_keep]
         input = input[, !(colnames(input) == "fraction_keep"), with = FALSE]
     }
@@ -284,9 +284,19 @@
 #' highest mean intensity for a given feature
 #' @keywords internal
 .getCorrectFraction = function(input) {
-    Intensity = Run = Fraction = NULL
-    
-    measurement_count = input[!is.na(Intensity) & Intensity > 0, 
+    Intensity = Run = Fraction = IsotopeLabelType = NULL
+
+    # For protein turnover: select fraction based on non-heavy peptide intensity
+    # only. Heavy peptides inherit the same fraction via the shared feature key.
+    # NA IsotopeLabelType represents peptides whose sequence cannot carry the
+    # label; these are treated the same as light for fraction selection purposes.
+    non_heavy_rows = input[(IsotopeLabelType == "L" | is.na(IsotopeLabelType)) &
+                               !is.na(Intensity) & Intensity > 0]
+    if (nrow(non_heavy_rows) > 0) {
+        input = non_heavy_rows
+    }
+
+    measurement_count = input[!is.na(Intensity) & Intensity > 0,
                               list(n_obs = data.table::uniqueN(Run)),
                               by = "Fraction"]
     which_max_measurements = which(measurement_count$n_obs == max(measurement_count$n_obs))