Merge pull request #553 from Sage-Bionetworks/GEN_1067_warn_for_uncha…

…nged_mutations

[GEN-1067] add functionality to warn for identical ref and tsa2

genie_registry/maf.py

@@ -11,9 +11,9 @@
 logger = logging.getLogger(__name__)
 
 
-def _check_tsa1_tsa2(df):
+def _check_allele_col_validity(df):
     """If maf file has both TSA1 and TSA2,
-    TSA1 must equal REF, or TSA1 must equal TSA2.
+    TSA1 must equal REF, or TSA1 must equal TSA2, and REF must not equal TSA2
     """
     tsa2_col_exist = process_functions.checkColExist(df, "TUMOR_SEQ_ALLELE2")
     tsa1_col_exist = process_functions.checkColExist(df, "TUMOR_SEQ_ALLELE1")
@@ -29,6 +29,16 @@ def _check_tsa1_tsa2(df):
                 "All values in TUMOR_SEQ_ALLELE1 must match all values in "
                 "REFERENCE_ALLELE or all values in TUMOR_SEQ_ALLELE2.\n"
             )
+    if (
+        tsa2_col_exist
+        and ref_col_exist
+        and not df.query("REFERENCE_ALLELE == TUMOR_SEQ_ALLELE2").empty
+    ):
+        error = (
+            f"{error}maf: Contains instances where values in REFERENCE_ALLELE match values in TUMOR_SEQ_ALLELE2. "
+            "This is invalid. Please correct.\n"
+        )
+        row_index = df.query("REFERENCE_ALLELE == TUMOR_SEQ_ALLELE2").index.values
     return error
 
 
@@ -260,7 +270,7 @@ def _validate(self, mutationDF):
         #             "start with 'chr' or any 'WT' values.\n"
         #         )
 
-        error = _check_tsa1_tsa2(mutationDF)
+        error = _check_allele_col_validity(mutationDF)
         total_error.write(error)
 
         if process_functions.checkColExist(mutationDF, "TUMOR_SAMPLE_BARCODE"):

tests/test_maf.py

@@ -1,3 +1,4 @@
+from cmath import nan
 from unittest.mock import mock_open, patch
 
 import pandas as pd
@@ -81,7 +82,7 @@ def test_firstcolumn_validation(maf_class):
             "N_DEPTH": [1, 2, 3, 4, 3],
             "N_REF_COUNT": [1, 2, 3, 4, 3],
             "N_ALT_COUNT": [1, 2, 3, 4, 3],
-            "TUMOR_SEQ_ALLELE2": ["A", "A", "A", "A", "A"],
+            "TUMOR_SEQ_ALLELE2": ["T", "A", "A", "A", "A"],
         }
     )
     order = [
@@ -196,7 +197,7 @@ def test_invalid_validation(maf_class):
     )
 
     with patch.object(
-        genie_registry.maf, "_check_tsa1_tsa2", return_value=""
+        genie_registry.maf, "_check_allele_col_validity", return_value=""
     ) as check_tsa1_tsa2:
         error, warning = maf_class._validate(mafDf)
         check_tsa1_tsa2.assert_called_once_with(mafDf)
@@ -241,47 +242,124 @@ def test_error__check_allele_col():
     assert warning == ""
 
 
-def test_invalid__check_tsa1_tsa2():
-    """Test the scenario in which maf file has TSA1 and TSA2 and fails"""
-    df = pd.DataFrame(
-        dict(
-            REFERENCE_ALLELE=["A", "A", "A"],
-            TUMOR_SEQ_ALLELE1=["B", "B", "B"],
-            TUMOR_SEQ_ALLELE2=["C", "C", "C"],
-        )
-    )
-    error = genie_registry.maf._check_tsa1_tsa2(df)
-    assert error == (
-        "maf: Contains both "
-        "TUMOR_SEQ_ALLELE1 and TUMOR_SEQ_ALLELE2 columns. "
-        "All values in TUMOR_SEQ_ALLELE1 must match all values in "
-        "REFERENCE_ALLELE or all values in TUMOR_SEQ_ALLELE2.\n"
-    )
-
-
 @pytest.mark.parametrize(
-    "df",
+    "test_df,expected_error",
     [
-        pd.DataFrame(
-            dict(
-                REFERENCE_ALLELE=["A", "A", "A"],
-                TUMOR_SEQ_ALLELE1=["C", "C", "C"],
-                TUMOR_SEQ_ALLELE2=["C", "C", "C"],
-            )
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE1=["C", "C", "C"],
+                    TUMOR_SEQ_ALLELE2=["C", "C", "C"],
+                )
+            ),
+            "",
         ),
-        pd.DataFrame(
-            dict(
-                REFERENCE_ALLELE=["C", "C", "C"],
-                TUMOR_SEQ_ALLELE1=["C", "C", "C"],
-                TUMOR_SEQ_ALLELE2=["A", "A", "A"],
-            )
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["C", "C", "C"],
+                    TUMOR_SEQ_ALLELE1=["C", "C", "C"],
+                    TUMOR_SEQ_ALLELE2=["A", "A", "A"],
+                )
+            ),
+            "",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE1=["B", "B", "B"],
+                    TUMOR_SEQ_ALLELE2=["C", "C", "C"],
+                )
+            ),
+            "maf: Contains both "
+            "TUMOR_SEQ_ALLELE1 and TUMOR_SEQ_ALLELE2 columns. "
+            "All values in TUMOR_SEQ_ALLELE1 must match all values in "
+            "REFERENCE_ALLELE or all values in TUMOR_SEQ_ALLELE2.\n",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE1=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE2=["A", "C", "C"],
+                )
+            ),
+            "maf: Contains instances where values in REFERENCE_ALLELE match values in TUMOR_SEQ_ALLELE2. "
+            "This is invalid. Please correct.\n",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE2=["A", "C", "C"],
+                )
+            ),
+            "maf: Contains instances where values in REFERENCE_ALLELE match values in TUMOR_SEQ_ALLELE2. "
+            "This is invalid. Please correct.\n",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE2=["C", "C", "C"],
+                )
+            ),
+            "",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    TUMOR_SEQ_ALLELE1=["C", "C", "C"],
+                )
+            ),
+            "",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=["A", "A", "A"],
+                    TUMOR_SEQ_ALLELE1=["B", "B", "B"],
+                    TUMOR_SEQ_ALLELE2=["A", "C", "C"],
+                )
+            ),
+            "maf: Contains both "
+            "TUMOR_SEQ_ALLELE1 and TUMOR_SEQ_ALLELE2 columns. "
+            "All values in TUMOR_SEQ_ALLELE1 must match all values in "
+            "REFERENCE_ALLELE or all values in TUMOR_SEQ_ALLELE2.\n"
+            "maf: Contains instances where values in REFERENCE_ALLELE match values in TUMOR_SEQ_ALLELE2. "
+            "This is invalid. Please correct.\n",
+        ),
+        (
+            pd.DataFrame(
+                dict(
+                    REFERENCE_ALLELE=[nan, "A", "A"],
+                    TUMOR_SEQ_ALLELE1=["B", nan, "B"],
+                    TUMOR_SEQ_ALLELE2=[nan, "C", "C"],
+                )
+            ),
+            "maf: Contains both "
+            "TUMOR_SEQ_ALLELE1 and TUMOR_SEQ_ALLELE2 columns. "
+            "All values in TUMOR_SEQ_ALLELE1 must match all values in "
+            "REFERENCE_ALLELE or all values in TUMOR_SEQ_ALLELE2.\n",
         ),
     ],
+    ids=[
+        "matching_tsa1_tsa2",
+        "matching_tsa1_ref",
+        "invalid_tsa1",
+        "identical_ref_tsa2",
+        "identical_ref_tsa2_missing_tsa1",
+        "valid_ref_tsa2_missing_tsa1",
+        "missing_tsa2_ref",
+        "invalid_tsa1_identical_ref_tsa2",
+        "NAs_in_allele_cole",
+    ],
 )
-def test_valid__check_tsa1_tsa2(df):
-    """Test valid TSA1 and TSA2"""
-    error = genie_registry.maf._check_tsa1_tsa2(df)
-    assert error == ""
+def test__check_allele_col_validity(test_df, expected_error):
+    error = genie_registry.maf._check_allele_col_validity(test_df)
+    assert error == expected_error
 
 
 def test_that__cross_validate_does_not_read_files_if_no_clinical_files(maf_class):