Merge pull request #21 from akikuno:develop-v1.0.3

Develop-v1.0.3

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "cstag"
-version = "1.0.2"
+version = "1.0.3"
 description = "Python module to manipulate the minimap2's CS tag"
 authors = ["Akihiro Kuno <[email protected]>"]
 homepage = "https://github.com/akikuno/cstag"

src/cstag/to_vcf.py

@@ -194,7 +194,7 @@ def replace_mutation_to_atmark(cs_tags: str) -> str:
     return "".join(cs if cs in {"A", "C", "G", "T"} else "@" for cs in cs_tags)
 
 
-def call_reference_depth(variant_annotations, cs_tags_list, positions_list) -> dict[int, int]:
+def call_reference_depth(variant_annotations, cs_tags_list, positions_list) -> dict[tuple(str, int), int]:
     cs_tags_normalized_length = normalize_read_lengths(cs_tags_list, positions_list)
     cs_replaced = [replace_mutation_to_atmark(cs_tags) for cs_tags in cs_tags_normalized_length]
 
@@ -204,19 +204,21 @@ def call_reference_depth(variant_annotations, cs_tags_list, positions_list) -> d
         v_idx = v.pos - min(positions_list)
         for cs in cs_replaced:
             if v.ref == cs[v_idx : v_idx + len(v.ref)]:
-                reference_depth[v.pos] += 1
+                reference_depth[(v.ref, v.pos)] += 1
 
     return dict(reference_depth)
 
 
-def add_vcf_fields(variant_annotations: list[Vcf], chrom: str, reference_depth: dict[int, int]) -> list[Vcf]:
+def add_vcf_fields(
+    variant_annotations: list[Vcf], chrom: str, reference_depth: dict[tuple(str, int), int]
+) -> list[Vcf]:
     """Add Chrom and VCF info (AD, RD, DP, and VAF) to immutable Vcf dataclass"""
     variant_counter = Counter((v.pos, v.ref, v.alt) for v in variant_annotations)
 
     updated_annotations = []
     for v in set(variant_annotations):
         ad = variant_counter[(v.pos, v.ref, v.alt)]
-        rd = reference_depth.get(v.pos, 0)
+        rd = reference_depth.get((v.ref, v.pos), 0)
         dp = rd + ad
         vaf = round(ad / dp, 3) if dp else 0
 

tests/test_to_vcf.py

@@ -173,7 +173,7 @@ def test_call_reference_depth():
     ]
     cs_tags_list = ["=A*cg=G+aa=T-ac=G", "*cg=G+aa=T-ac=G", "=G-t=ACG"]
     positions_list = [10, 11, 12]
-    expected_output = {12: 1}
+    expected_output = {("G", 12): 1}
 
     result = call_reference_depth(variant_annotations, cs_tags_list, positions_list)
     assert result == expected_output, f"Expected {expected_output}, but got {result}"
@@ -186,7 +186,7 @@ def test_add_vcf_fields():
         Vcf(chrom=None, pos=2, ref="G", alt="C", info=VcfInfo(dp=None, rd=None, ad=None, vaf=None)),
     ]
     sample_chrom = "chr1"
-    sample_reference_depth = {1: 10, 2: 5}
+    sample_reference_depth = {("A", 1): 10, ("G", 2): 5}
     result = add_vcf_fields(sample_variant_annotations, sample_chrom, sample_reference_depth)
     result = sorted(result, key=lambda x: (chrom_sort_key(x.chrom), x.pos))
     assert result[0].chrom == "chr1"