Merge pull request #65 from akikuno/patch-v0.5.5

Patch v0.5.5 (Version Bump to v0.5.5.1)

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "DAJIN2"
-version = "0.5.5"
+version = "0.5.5.1"
 description = "One-step genotyping tools for targeted long-read sequencing"
 authors = ["Akihiro Kuno <[email protected]>"]
 readme = "README.md"

src/DAJIN2/core/core.py

@@ -123,24 +123,22 @@ def execute_sample(arguments: dict):
     preprocess.cache_mutation_loci(ARGS, is_control=False)
 
     # ============================================================
-    # Detect and mapping SV (insertion/inversion) alleles
+    # Detect and mapping SV alleles
     # ============================================================
     paths_predefined_fasta: set[str] = {
         str(Path(ARGS.tempdir, ARGS.sample_name, "fasta", f"{allele}.fasta")) for allele in ARGS.fasta_alleles.keys()
     }
 
     preprocess.detect_insertions(ARGS.tempdir, ARGS.sample_name, ARGS.control_name, ARGS.fasta_alleles)
-    preprocess.detect_inversions(ARGS.tempdir, ARGS.sample_name, ARGS.control_name, ARGS.fasta_alleles)
 
     paths_sv_fasta = {str(p) for p in Path(ARGS.tempdir, ARGS.sample_name, "fasta").glob("insertion*.fasta")}
-    paths_sv_fasta |= {str(p) for p in Path(ARGS.tempdir, ARGS.sample_name, "fasta").glob("inversion*.fasta")}
     paths_sv_fasta -= paths_predefined_fasta
 
     if paths_sv_fasta:
-        # mapping to SV (insertion/inversion) alleles
+        # mapping to SV alleles
         preprocess.generate_sam(ARGS, paths_sv_fasta, is_control=True, is_sv=True)
         preprocess.generate_sam(ARGS, paths_sv_fasta, is_control=False, is_sv=True)
-        # add SV (insertion/inversion) alleles to ARGS.fasta_alleles
+        # add SV alleles to ARGS.fasta_alleles
         for path_fasta_sv in paths_sv_fasta:
             allele, seq = Path(path_fasta_sv).read_text().strip().split("\n")
             allele = allele.replace(">", "")

src/DAJIN2/core/preprocess/sv_handler.py

@@ -45,9 +45,7 @@ def _check_duplicates_of_sets(set1: set[str], set2: set[str]) -> bool:
     return len(union_set) != total_elements
 
 
-def add_unique_allele_keys(
-    fasta_sv_alleles: dict[str, str], FASTA_ALLELES: dict[str, set], key: str
-) -> dict[str, str]:
+def add_unique_allele_keys(fasta_sv_alleles: dict[str, str], FASTA_ALLELES: dict[str, set], key: str) -> dict[str, str]:
     """
     Update keys to avoid duplicating user-specified alleles.
     If the allele 'insertion1' exists in FASTA_ALLELES, increment the digits.
@@ -56,7 +54,7 @@ def add_unique_allele_keys(
     user_defined_alleles = set(FASTA_ALLELES)
     key_duplicated_alleles = {allele for allele in user_defined_alleles if key in allele}
     if key_duplicated_alleles == set():
-        return {f"inversion{i+1}": value for i, value in enumerate(fasta_sv_alleles.values())}
+        return {f"{key}{i+1}": value for i, value in enumerate(fasta_sv_alleles.values())}
 
     key_candidate_alleles = set()
     digits = 2

src/DAJIN2/utils/config.py

@@ -8,7 +8,7 @@
 
 from sklearn.exceptions import ConvergenceWarning
 
-DAJIN_VERSION = "0.5.5"
+DAJIN_VERSION = "0.5.5.1"
 DAJIN_RESULTS_DIR = Path("DAJIN_Results")
 TEMP_ROOT_DIR = Path(DAJIN_RESULTS_DIR, ".tempdir")
 

src/DAJIN2/utils/cssplits_handler.py

@@ -26,6 +26,7 @@ def find_n_boundaries(cssplits: list[str]) -> tuple[int, int]:
 
     return left_idx_n - 1, right_idx_n + 1
 
+
 # ToDo: 点変異が扱えていない。convert_cssplits_to_cstagをinversion検出用に改変すれば、必要なくなる。
 def convert_cssplit_to_dna(cssplit: str) -> str:
     """Convert a cssplit to a plain DNA sequence."""