RNApolis 0.4.2__py3-none-any.whl → 0.4.4__py3-none-any.whl

{RNApolis-0.4.2.dist-info → RNApolis-0.4.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: RNApolis
-Version: 0.4.2
+Version: 0.4.4
 Summary: A Python library containing RNA-related bioinformatics functions and classes
 Home-page: https://github.com/tzok/rnapolis-py
 Author: Tomasz Zok

{RNApolis-0.4.2.dist-info → RNApolis-0.4.4.dist-info}/RECORD

@@ -6,12 +6,12 @@ rnapolis/molecule_filter.py,sha256=hB6-nXgjmw7FAsQ3bj0cZ2FvuW2I1PXunEfcdwEUB1o,7
 rnapolis/motif_extractor.py,sha256=duHvpi9Ulcny9K60E6VBpz5RpJZw-KdTB4_Ph0iP478,774
 rnapolis/parser.py,sha256=wCA9rXqt51iLECgeBqOShFpuT8JwanNkHYD5uXYvLzU,13988
 rnapolis/rfam_folder.py,sha256=SjiiyML_T1__saruFwSMJEoQ7Y55GIU8ktS8ZUn5-fw,11111
-rnapolis/tertiary.py,sha256=ya7A2HyhPZSbA4rC8E43-mRIhR73-AeBuLE2GRpOTxk,20347
+rnapolis/tertiary.py,sha256=SQyiYWA0RJhAK70f88CKZvS4EzGKHQ2RoL1s4MueEDQ,21657
 rnapolis/transformer.py,sha256=V9nOQvdq4-p7yUWo0vQg0CDQMpmyxz9t4TMSRVEKHnw,1817
 rnapolis/util.py,sha256=IdquFO3PV1_KDqodjupzm0Rqvgy0CeSzxGHaGEHYXVU,543
-RNApolis-0.4.2.dist-info/LICENSE,sha256=ZGRu12MzCgbYA-Lt8MyBlmjvPZh7xfiD5u5wBx0enq4,1066
-RNApolis-0.4.2.dist-info/METADATA,sha256=1rZv7syiOsppxC9PAy0kYmgBXABKCX4gWZaqdtpbvis,54322
-RNApolis-0.4.2.dist-info/WHEEL,sha256=HiCZjzuy6Dw0hdX5R3LCFPDmFS4BWl8H-8W39XfmgX4,91
-RNApolis-0.4.2.dist-info/entry_points.txt,sha256=foN2Pn5e-OzEz0fFmNoX6PnFSZFQntOlY8LbognP5F0,308
-RNApolis-0.4.2.dist-info/top_level.txt,sha256=LcO18koxZcWoJ21KDRRRo_tyIbmXL5z61dPitZpy8yc,9
-RNApolis-0.4.2.dist-info/RECORD,,
+RNApolis-0.4.4.dist-info/LICENSE,sha256=ZGRu12MzCgbYA-Lt8MyBlmjvPZh7xfiD5u5wBx0enq4,1066
+RNApolis-0.4.4.dist-info/METADATA,sha256=irtWJbeg1LWun2r3WtnsnDDSHlLvru0hO9wz1e67cIE,54322
+RNApolis-0.4.4.dist-info/WHEEL,sha256=cVxcB9AmuTcXqmwrtPhNK88dr7IR_b6qagTj0UvIEbY,91
+RNApolis-0.4.4.dist-info/entry_points.txt,sha256=foN2Pn5e-OzEz0fFmNoX6PnFSZFQntOlY8LbognP5F0,308
+RNApolis-0.4.4.dist-info/top_level.txt,sha256=LcO18koxZcWoJ21KDRRRo_tyIbmXL5z61dPitZpy8yc,9
+RNApolis-0.4.4.dist-info/RECORD,,

{RNApolis-0.4.2.dist-info → RNApolis-0.4.4.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (72.2.0)
+Generator: setuptools (74.1.2)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

rnapolis/tertiary.py

@@ -7,6 +7,7 @@ from typing import Dict, List, Optional, Set, Tuple, Union
 
 import numpy
 import numpy.typing
+
 from rnapolis.common import (
     BasePair,
     BpSeq,
@@ -16,6 +17,7 @@ from rnapolis.common import (
     Residue,
     ResidueAuth,
     ResidueLabel,
+    Saenger,
     Stacking,
 )
 
@@ -497,34 +499,71 @@ class Mapping2D3D:
 
     @cached_property
     def strands_sequences(self) -> List[Tuple[str, str]]:
-        result = defaultdict(list)
+        nucleotides = list(filter(lambda r: r.is_nucleotide, self.structure3d.residues))
 
-        for i, residue in enumerate(
-            filter(lambda r: r.is_nucleotide, self.structure3d.residues)
-        ):
-            if i > 0 and self.find_gaps:
-                previous = self.structure3d.residues[i - 1]
+        if not nucleotides:
+            return []
 
-                if (
-                    not previous.is_connected(residue)
-                    and previous.chain == residue.chain
-                ):
-                    for k in range(residue.number - previous.number - 1):
-                        result[residue.chain].append("?")
+        result = [(nucleotides[0].chain, [nucleotides[0].one_letter_name])]
 
-            result[residue.chain].append(residue.one_letter_name)
+        for i in range(1, len(nucleotides)):
+            previous = nucleotides[i - 1]
+            residue = nucleotides[i]
 
-        return [(chain, "".join(sequence)) for chain, sequence in result.items()]
+            if residue.chain != previous.chain:
+                result.append((residue.chain, [residue.one_letter_name]))
+            else:
+                if self.find_gaps:
+                    if not previous.is_connected(residue):
+                        for k in range(residue.number - previous.number - 1):
+                            result[-1][1].append("?")
+                result[-1][1].append(residue.one_letter_name)
+
+        return [(chain, "".join(sequence)) for chain, sequence in result]
 
     @cached_property
     def bpseq(self) -> BpSeq:
-        return self.__generate_bpseq(
-            [
-                base_pair
-                for base_pair in self.base_pairs
-                if base_pair.is_canonical and base_pair.nt1 < base_pair.nt2
-            ]
-        )
+        def pair_scoring_function(pair: BasePair3D) -> int:
+            if pair.saenger is not None:
+                if pair.saenger in (Saenger.XIX, Saenger.XX):
+                    return 0, pair.nt1, pair.nt2
+                else:
+                    return 1, pair.nt1, pair.nt2
+
+            sequence = "".join(
+                sorted(
+                    [
+                        pair.nt1_3d.one_letter_name.upper(),
+                        pair.nt2_3d.one_letter_name.upper(),
+                    ]
+                )
+            )
+            if sequence in ("AU", "AT", "CG"):
+                return 0, pair.nt1, pair.nt2
+            return 1, pair.nt1, pair.nt2
+
+        canonical = [
+            base_pair
+            for base_pair in self.base_pairs
+            if base_pair.is_canonical and base_pair.nt1 < base_pair.nt2
+        ]
+
+        while True:
+            matches = defaultdict(set)
+
+            for base_pair in canonical:
+                matches[base_pair.nt1_3d].add(base_pair)
+                matches[base_pair.nt2_3d].add(base_pair)
+
+            for pairs in matches.values():
+                if len(pairs) > 1:
+                    pairs = sorted(pairs, key=pair_scoring_function)
+                    canonical.remove(pairs[-1])
+                    break
+            else:
+                break
+
+        return self.__generate_bpseq(canonical)
 
     def __generate_bpseq(self, base_pairs):
         result: Dict[int, List] = {}