Changed alignment library

bmlsa/aligner.py

@@ -1,31 +1,41 @@
-from skbio.alignment import StripedSmithWaterman
+from Bio.Align import PairwiseAligner, substitution_matrices
+from exceptions import UnexpectedAlignmentResult
 
-from datatypes import Annotation
+from datatypes import AlignedSequence
 
 
-def generate_annotated_positions(sequence: str, annotations: dict[str, Annotation]):
+def protein_align_many_to_one_ssw(sequence: str, queries: dict[str, AlignedSequence]):
     annotation_pairs = {}
-    align = StripedSmithWaterman(sequence)
+    aligner = PairwiseAligner(one_alignment_only=True)
-    for id, annot in annotations.items():
+    aligner.mode = "local"
-        alignment = align(annot.sequence)
+    aligner.substitution_matrix = substitution_matrices.load("BLOSUM62")
-        score, aligned_start, aligned_end = (
+    aligner.extend_gap_score = -1
-            alignment.optimal_alignment_score,
+    aligner.open_gap_score = -11
-            alignment.query_begin,
+    for id, query in queries.items():
-            alignment.query_end,
+        try:
+            alignments = aligner.align(sequence, query.sequence)
+        except ValueError:
+            continue
+        if len(alignments) > 1:
+            raise UnexpectedAlignmentResult(
+                "More than one alignment resulted from a single query."
             )
-        annotation_pairs[id] = Annotation(
+        for alignment in alignments:
+            score, query_aligned = (alignment.score, alignment.aligned[0][0])
+            aligned_start, aligned_end = query_aligned
+            annotation_pairs[id] = AlignedSequence(
                 id,
-            annot.sequence,
+                query.sequence,
-            annot.name,
+                query.name,
-            annot.description,
+                query.description,
-            annot.start,
+                query.start,
-            annot.end,
+                query.end,
-            annot.score,
+                query.score,
-        ), Annotation(
+            ), AlignedSequence(
                 id,
-            alignment.aligned_target_sequence,
+                alignment.query,
-            annot.name,
+                query.name,
-            annot.description,
+                query.description,
                 aligned_start,
                 aligned_end,
                 score,

bmlsa/cli.py

@@ -1,7 +1,7 @@
 import os
 import argparse
 from Bio import SeqIO
-from aligner import generate_annotated_positions
+from aligner import protein_align_many_to_one_ssw
 
 from persistence import read_annotations_from_csv, save_alignments_to_csv
 
@@ -78,7 +78,7 @@ def main():
     )
     with open(args.sequence, "r") as sequence_fd:
         for sequence in SeqIO.parse(sequence_fd, "fasta"):
-            aligned_annotations = generate_annotated_positions(
+            aligned_annotations = protein_align_many_to_one_ssw(
                 str(sequence.seq), given_annotations
             )
             save_alignments_to_csv(

bmlsa/datatypes.py

@@ -1,4 +1,4 @@
-class Annotation:
+class AlignedSequence:
     def __init__(
         self,
         id: str,

bmlsa/exceptions.py

@@ -0,0 +1,2 @@
+class UnexpectedAlignmentResult(Exception):
+    pass

bmlsa/persistence.py

@@ -1,5 +1,5 @@
 import csv
-from datatypes import Annotation
+from datatypes import AlignedSequence
 
 
 def read_annotations_from_csv(
@@ -37,7 +37,7 @@ def read_annotations_from_csv(
             start = row[start_ind] if start_header else None
             end = row[end_ind] if end_header else None
             sequence = row[sequence_ind]
-            annotations[id] = Annotation(
+            annotations[id] = AlignedSequence(
                 id,
                 sequence,
                 name,
@@ -49,7 +49,7 @@ def read_annotations_from_csv(
 
 
 def save_alignments_to_csv(
-    aligned_pairs: dict[str, tuple[Annotation, Annotation]], output_path: str
+    aligned_pairs: dict[str, tuple[AlignedSequence, AlignedSequence]], output_path: str
 ):
     with open(output_path, "w") as output_fd:
         writer = csv.writer(output_fd)