Merge pull request #38 from broadinstitute/hm-fix-long-fragment-length

Fix --cluster-from-fragments bug with long fragment length
broadinstitute · Jun 12, 2020 · 8d42266 · 8d42266
2 parents 8c86806 + ad5e062
commit 8d42266
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/catch/genome.py b/catch/genome.py
@@ -83,7 +83,7 @@ def fragments(seq):
                 if include_full_end and len(fragment) < fragment_length:
                     # This is at the end of seq; instead, use the last
                     # fragment_length nt
-                    yield seq[(len(seq) - fragment_length):]
+                    yield seq[max(0, len(seq) - fragment_length):]
                 else:
                     yield fragment
 

diff --git a/catch/tests/test_genome.py b/catch/tests/test_genome.py
@@ -92,6 +92,13 @@ def test_break_into_fragments_from_one_seq_with_full_end(self):
                 OrderedDict([('0', 'ATC'), ('1', 'GTT'), ('2', 'TAA')]))
         self.assertEqual(broken, expected_genome)
 
+    def test_break_into_fragments_from_one_seq_with_full_end_and_long_fragment(self):
+        genome_one = genome.Genome.from_one_seq('ATCGTTAA')
+        broken = genome_one.break_into_fragments(12, include_full_end=True)
+        expected_genome = genome.Genome.from_chrs(
+                OrderedDict([('0', 'ATCGTTAA')]))
+        self.assertEqual(broken, expected_genome)
+
     def test_break_into_fragments_from_chrs(self):
         genome_two_chrs = genome.Genome.from_chrs(
                 OrderedDict([("chr1", 'ATCGTTAA'), ("chr2", 'AATTCCGGG')]))