Changes yet to be tested: removes outliers from reference calls too

broadinstitute · Oct 28, 2024 · 40248ae · 40248ae
1 parent 9d763a5
commit 40248ae
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 12 deletions.
diff --git a/src/RdTest/RdTest.R b/src/RdTest/RdTest.R
@@ -1183,12 +1183,11 @@ runRdTest<-function(bed)
   #Exclude outlier samples only if non-outlier samples exist
   if (!is.null(opt$outlierSampleIds)) {
     outlier_ids <- readLines(opt$outlierSampleIds)
-    outlier_samples <- intersect(sampleIDs, outlier_ids)
     non_outlier_samples <- setdiff(sampleIDs, outlier_ids)
-    if (length(outlier_samples) > 0 && length(non_outlier_samples) > 0) {
+    if (length(non_outlier_samples) > 0) {
       sampleIDs <- non_outlier_samples
-      cnv_matrix <- cnv_matrix[!(rownames(cnv_matrix) %in% outlier_samples), , drop = FALSE]
     }
+    cnv_matrix <- cnv_matrix[!(rownames(cnv_matrix) %in% outlier_ids), , drop = FALSE]
   }
   samplesPrior <-unlist(strsplit(as.character(sampleIDs),split=","))
 

diff --git a/src/svtk/svtk/cli/baf_test.py b/src/svtk/svtk/cli/baf_test.py
@@ -53,14 +53,14 @@ def preprocess(chrom, start, end, tbx, samples, window=None, called_samples=None
         return bafs, bafs, called_samples
     bafs.columns = ['chr', 'pos', 'baf', 'sample']
 
-    # Exclude outlier samples only if non-outlier samples exist
-    filtered_samples = []
-    retained_samples = called_samples
-    non_outlier_samples = [s for s in called_samples if s not in outlier_sample_ids]
-    if (non_outlier_samples):
-        filtered_samples = [s for s in called_samples if s in outlier_sample_ids]
-        retained_samples = non_outlier_samples
-    bafs = bafs[bafs['sample'].isin(samples) & ~bafs['sample'].isin(filtered_samples)]
+    bafs = bafs[bafs['sample'].isin(samples) & ~bafs['sample'].isin(outlier_sample_ids)]
+
+    # Exclude outlier samples from called samples only if non-outlier samples exist
+    non_outlier_called_samples = [s for s in called_samples if s not in outlier_sample_ids]
+    if non_outlier_called_samples:
+        retained_samples = non_outlier_called_samples
+    else:
+        retained_samples = called_samples
 
     if bafs.empty:
         return bafs, bafs, retained_samples

diff --git a/src/svtk/svtk/pesr/pesr_test.py b/src/svtk/svtk/pesr/pesr_test.py
@@ -140,7 +140,7 @@ def test_record(self, record):
     def choose_background(self, record, whitelist=None, blacklist=None):
         # Select called and background samples
         called = svu.get_called_samples(record)
-        background = [s for s in self.samples if s not in called]
+        background = [s for s in self.samples if s not in called and s not in self.outlier_sample_ids]
 
         # Exclude outlier samples only if non-outlier samples exist
         non_outlier_called = [s for s in called if s not in self.outlier_sample_ids]