diff --git a/src/naibr/__main__.py b/src/naibr/__main__.py
index c3eed80..52b8de2 100644
--- a/src/naibr/__main__.py
+++ b/src/naibr/__main__.py
@@ -58,6 +58,7 @@
     input_candidates,
     is_proper_chrom,
     parallel_execute,
+    UnionDict,
     write_novel_adjacencies,
 )
 
@@ -286,27 +287,6 @@ def parse_args(args):
     return file_configs
 
 
-class UnionDict(collections.defaultdict):
-    """Extends defaultdict to allow combining with other UnionDicts or defaultdicts"""
-
-    def combine(self, other):
-        if not isinstance(other, type(self)) and not isinstance(other, collections.defaultdict):
-            raise TypeError(f"Can only combine with other UnionDicts or defaultdicts ({type(other)})")
-
-        if self.default_factory is not other.default_factory:
-            raise ValueError("Can only combine UnionDicts with the same default_factory")
-
-        if self.default_factory is list:
-            for k, v in other.items():
-                self[k].extend(v)
-
-        elif self.default_factory is set:
-            for k, v in other.items():
-                self[k].update(v)
-        else:
-            raise ValueError("Can only combine UnionDicts with default_factory of list or set")
-
-
 def run(configs):
     starttime = time.time()
 
diff --git a/src/naibr/utils.py b/src/naibr/utils.py
index 1d2ea45..8fd736b 100644
--- a/src/naibr/utils.py
+++ b/src/naibr/utils.py
@@ -606,3 +606,24 @@ def parse_naibr(lines):
     cands = [c for c in cands if abs(c[1] - c[3]) > min_sv or c[0] != c[2]]
     logger.info(f"Found {n_total:,} candidates in file of which {len(cands):,} are long enough")
     return cands
+
+
+class UnionDict(defaultdict):
+    """Extends defaultdict to allow combining with other UnionDicts or defaultdicts"""
+
+    def combine(self, other):
+        if not isinstance(other, type(self)) and not isinstance(other, defaultdict):
+            raise TypeError(f"Can only combine with other UnionDicts or defaultdicts ({type(other)})")
+
+        if self.default_factory is not other.default_factory:
+            raise ValueError("Can only combine UnionDicts with the same default_factory")
+
+        if self.default_factory is list:
+            for k, v in other.items():
+                self[k].extend(v)
+
+        elif self.default_factory is set:
+            for k, v in other.items():
+                self[k].update(v)
+        else:
+            raise ValueError("Can only combine UnionDicts with default_factory of list or set")
diff --git a/tests/test_utils.py b/tests/test_utils.py
index 029b64d..f3445e5 100644
--- a/tests/test_utils.py
+++ b/tests/test_utils.py
@@ -1,6 +1,8 @@
 from io import StringIO
 
-from naibr.utils import input_candidates
+import pytest
+
+from naibr.utils import input_candidates, UnionDict
 
 
 def test_input_candates_bepde_format():
@@ -40,3 +42,47 @@ def test_input_candates_filter_too_short():
     s += "chr1\t10000\t10000\tchr1\t14000\t14000\t-+\n"
     with StringIO(s) as f:
         assert len(input_candidates(f, min_sv=1000)) == 2
+
+
+def test_union_dict_combine_lists():
+    d1 = UnionDict(list)
+    d1["a"].extend([1, 2, 3])
+    d1["b"].extend([7, 8, 9])
+
+    d2 = UnionDict(list)
+    d2["a"].extend([3, 4, 5])
+    d2["b"].extend([8, 9, 10])
+
+    d1.combine(d2)
+    assert sorted(d1["a"]) == sorted([1, 2, 3, 3, 4, 5])
+    assert sorted(d1["b"]) == sorted([7, 8, 8, 9, 9, 10])
+
+
+def test_union_dict_combine_sets():
+    d1 = UnionDict(set)
+    d1["a"].union({1, 2, 3})
+    d1["b"].union({7, 8, 9})
+
+    d2 = UnionDict(set)
+    d2["a"].union({3, 4, 5})
+    d2["b"].union({8, 9, 10})
+
+    d1.combine(d2)
+    assert d1["a"] - {1, 2, 3, 4, 5} == set()
+    assert d1["b"] - {7, 8, 9, 10} == set()
+
+
+def test_union_dict_combine_different_fails():
+    d1 = UnionDict(set)
+    d1["a"].union({1, 2, 3})
+    d1["b"].union({7, 8, 9})
+
+    d2 = UnionDict(list)
+    d2["a"].extend([3, 4, 5])
+    d2["b"].extend([8, 9, 10])
+
+    with pytest.raises(ValueError):
+        d1.combine(d2)
+
+    with pytest.raises(ValueError):
+        d2.combine(d1)