fixed bug in inferring ancestral states with ambiguous states

YosefLab · Feb 2, 2024 · 6a600eb · 6a600eb
1 parent a9c8cbd
commit 6a600eb
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 10 deletions.
diff --git a/cassiopeia/data/utilities.py b/cassiopeia/data/utilities.py
@@ -61,17 +61,27 @@ def get_lca_characters(
             all_states = [
                 vec[i] for vec in vecs if vec[i] != missing_state_indicator
             ]
-            chars = set.intersection(
-                *map(
-                    set,
-                    [
-                        state if is_ambiguous_state(state) else [state]
-                        for state in all_states
-                    ],
+
+            # this check is specifically if all_states consists of a single
+            # ambiguous state.
+            if len(all_states) == 1:
+                state = all_states[0]
+                if is_ambiguous_state(state) and len(state) == 1:
+                    lca_vec[i] = state[0]
+                else:
+                    lca_vec[i] = all_states[0]
+            else:
+                chars = set.intersection(
+                    *map(
+                        set,
+                        [
+                            state if is_ambiguous_state(state) else [state]
+                            for state in all_states
+                        ],
+                    )
                 )
-            )
-            if len(chars) == 1:
-                lca_vec[i] = list(chars)[0]
+                if len(chars) == 1:
+                    lca_vec[i] = list(chars)[0]
     return lca_vec
 
 

diff --git a/test/data_tests/data_utilities_test.py b/test/data_tests/data_utilities_test.py
@@ -316,6 +316,17 @@ def test_lca_characters_ambiguous(self):
         )
         self.assertEqual(ret_vec, [1, 2, 3, 0, 5])
 
+    def test_lca_characters_ambiguous_and_missing(self):
+        vecs = [
+            [(1, 1), (0, 2), (3,), (4,), (5,)],
+            [1, -1, -1, 3, -1],
+            [1, -1, 3, 2, -1],
+        ]
+        ret_vec = data_utilities.get_lca_characters(
+            vecs, missing_state_indicator=-1
+        )
+        self.assertEqual(ret_vec, [1, (0,2), 3, 0, 5])
+
     def test_resolve_most_abundant(self):
         state = (1, 2, 3, 3)
         self.assertEqual(data_utilities.resolve_most_abundant(state), 3)