Merge pull request #132 from sudlab/ns-rse/90-zip-blocks

isoslam/all_introns_counts_and_info.py

@@ -150,8 +150,8 @@ def fragment_iterator(read_iterator):
             ## If not, we should cull assignment to those events
             ## We are comparing ret vs spliced, so instead of adding 1 to both
             ## we will just ignore those few
-            block_starts1, block_ends1 = zip(*read1.get_blocks())
-            block_starts2, block_ends2 = zip(*read2.get_blocks())
+            block_starts1, block_ends1 = isoslam.zip_blocks(read1)
+            block_starts2, block_ends2 = isoslam.zip_blocks(read2)
             # RETAINED
             read1_within_intron = {}
 

isoslam/isoslam.py

@@ -1,7 +1,7 @@
 """IsoSLAM module."""
 
 from collections import defaultdict
-from collections.abc import Generator
+from collections.abc import Generator, Iterator
 from pathlib import Path
 from typing import Any
 
@@ -179,3 +179,20 @@ def extract_utron(features: dict[str, Any], gene_transcript: Any, coordinates: A
         untranslated_regions = [coordinates[transcript] for transcript in gene_transcript[features["transcript"]]]
         return sum(untranslated_regions, [])
     return []
+
+
+def zip_blocks(read: AlignedSegment) -> Iterator[tuple[Any, ...]]:
+    """
+    Zip the block starts and ends into two lists.
+
+    Parameters
+    ----------
+    read : AlignedSegment
+        An individual aligned segment read from a ''.bam'' file.
+
+    Returns
+    -------
+    tuple[list[int], list[int]]
+        Tuple of two lists of integers the first is start location, the second is the end location.
+    """
+    return zip(*read.get_blocks())

tests/test_isoslam.py

@@ -280,3 +280,44 @@ def test_extract_utron(
     assert len(untranslated_region) == length
     if len(untranslated_region):
         assert untranslated_region[0][3] == transcript_id  # type: ignore[misc]
+
+
+@pytest.mark.parametrize(
+    ("aligned_segment", "expected_start", "expected_end"),
+    [
+        pytest.param(  # type: ignore[misc]
+            "aligned_segment_unassigned_28584",
+            (28584, 28704),
+            (28704, 28733),
+            id="28584 - Assignment and Transcript are None",
+        ),
+        pytest.param(
+            "aligned_segment_assigned_17814",
+            (17814, 18027),
+            (17855, 18136),
+            id="17814 - Assigned to MSTRG.63147",
+        ),
+        pytest.param(
+            "aligned_segment_assigned_14770",
+            (14770,),
+            (14876,),
+            id="14770 - Assigned to MSTRG.63147",
+        ),
+        pytest.param(
+            "aligned_segment_assigned_15967",
+            (15967,),
+            (16117,),
+            id="15967 - Assigned to MSTRG.63147",
+        ),
+    ],
+)
+def test_zip_blocks(
+    aligned_segment: str,
+    expected_start: tuple[int],
+    expected_end: tuple[int],
+    request: pytest.FixtureRequest,
+) -> None:
+    """Test that block starts and ends are zipped correctly."""
+    start_blocks, end_blocks = isoslam.zip_blocks(request.getfixturevalue(aligned_segment))
+    assert start_blocks == expected_start
+    assert end_blocks == expected_end