Merged dev into main, resolved conflicts by accepting dev version

PyEED · Jun 4, 2024 · 5e52f48 · 5e52f48
2 parents 2e57aa4 + aec3a21
commit 5e52f48
Show file tree

Hide file tree

Showing 30 changed files with 351 additions and 4,920 deletions.
diff --git a/.DS_Store b/.DS_Store
diff --git a/docs/.DS_Store b/docs/.DS_Store
diff --git a/docs/examples/.DS_Store b/docs/examples/.DS_Store
diff --git a/docs/examples/image.png b/docs/examples/image.png
diff --git a/docs/examples/logs/py4cytoscape.log.1 b/docs/examples/logs/py4cytoscape.log.1
diff --git a/docs/examples/logs/py4cytoscape.log.2 b/docs/examples/logs/py4cytoscape.log.2
diff --git a/docs/examples/logs/py4cytoscape.log.3 b/docs/examples/logs/py4cytoscape.log.3
diff --git a/docs/examples/logs/py4cytoscape.log.4 b/docs/examples/logs/py4cytoscape.log.4
diff --git a/pyeed/core/abstractannotation.py b/pyeed/core/abstractannotation.py
@@ -49,7 +49,7 @@ class AbstractAnnotation(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/alignmentresult.py b/pyeed/core/alignmentresult.py
@@ -53,7 +53,7 @@ class AlignmentResult(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/annotation.py b/pyeed/core/annotation.py
@@ -10,3 +10,5 @@ class Annotation(Enum):
     FAMILY = "http://semanticscience.org/resource/SIO_001380"
     MOTIVE = "http://semanticscience.org/resource/SIO_000131"
     CODING_SEQ = "http://semanticscience.org/resource/SIO_001276"
+    ALPHAHELIX = "http://semanticscience.org/resource/SIO_010468"
+    BETASTRAND = "http://semanticscience.org/resource/SIO_010469"
diff --git a/pyeed/core/blastdata.py b/pyeed/core/blastdata.py
@@ -90,7 +90,7 @@ class BlastData(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/clustalomegaresult.py b/pyeed/core/clustalomegaresult.py
@@ -32,7 +32,7 @@ class ClustalOmegaResult(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/dnarecord.py b/pyeed/core/dnarecord.py
@@ -32,7 +32,7 @@ class DNARecord(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/organism.py b/pyeed/core/organism.py
@@ -108,7 +108,7 @@ class Organism(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/pairwisealignmentresult.py b/pyeed/core/pairwisealignmentresult.py
@@ -60,7 +60,7 @@ class PairwiseAlignmentResult(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/proteinrecord.py b/pyeed/core/proteinrecord.py
@@ -75,7 +75,7 @@ class ProteinRecord(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)
@@ -145,7 +145,6 @@ def get_id(cls, protein_id: str) -> "ProteinRecord":
 
         import nest_asyncio
 
-
         nest_asyncio.apply()
 
         if isinstance(protein_id, list) and all(isinstance(x, str) for x in protein_id):
@@ -166,7 +165,6 @@ def get_ids(cls, accession_ids: List[str]) -> List["ProteinRecord"]:
 
         import nest_asyncio
 
-
         nest_asyncio.apply()
 
         return asyncio.run(

diff --git a/pyeed/core/region.py b/pyeed/core/region.py
@@ -43,7 +43,7 @@ class Region(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/regionset.py b/pyeed/core/regionset.py
@@ -35,7 +35,7 @@ class RegionSet(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _object_terms: Set[str] = PrivateAttr(

diff --git a/pyeed/core/sequence.py b/pyeed/core/sequence.py
@@ -38,7 +38,7 @@ class Sequence(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/sequencerecord.py b/pyeed/core/sequencerecord.py
@@ -103,7 +103,7 @@ class SequenceRecord(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/site.py b/pyeed/core/site.py
@@ -35,7 +35,7 @@ class Site(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/core/standardnumbering.py b/pyeed/core/standardnumbering.py
@@ -47,7 +47,7 @@ class StandardNumbering(
 
     _repo: Optional[str] = PrivateAttr(default="https://github.com/PyEED/pyeed")
     _commit: Optional[str] = PrivateAttr(
-        default="63f43b11e0d359e1d0a1f541cea25dd484ad0072"
+        default="5ca1d8073b90b91effc0fe9e3aaa578caf05980f"
     )
 
     _raw_xml_data: Dict = PrivateAttr(default_factory=dict)

diff --git a/pyeed/fetch/pdbmapper.py b/pyeed/fetch/pdbmapper.py
@@ -1 +1,138 @@
+from __future__ import annotations
 
+import json
+import logging
+from typing import TYPE_CHECKING, List
+
+from pyeed.core import Annotation, Organism
+
+if TYPE_CHECKING:
+    from pyeed.core import ProteinRecord
+
+LOGGER = logging.getLogger(__name__)
+LOGGER.setLevel(logging.DEBUG)
+LOGGER.addHandler(logging.StreamHandler())
+
+
+class PDBMapper:
+    def __init__(self):
+        pass
+
+    def map_pdb_data(self, pdb_data: str) -> List[ProteinRecord]:
+        from pyeed.core import ProteinRecord
+
+        records = []
+
+        pdb_entries = json.loads(pdb_data)
+        entries = pdb_entries["data"]["entry"]["polymer_entities"]
+        try:
+            assert len(entries) > 0, f"No entries found in the PDB data {pdb_data}"
+        except TypeError:
+            return []
+        for entry in entries:
+            structure_id = entry["rcsb_id"]
+            seq_info = entry["rcsb_polymer_entity_container_identifiers"]
+            seq_organism = entry["rcsb_entity_source_organism"]
+            try:
+                seq_id = next(
+                    (
+                        identifier["database_accession"]
+                        for identifier in seq_info["reference_sequence_identifiers"]
+                        if "database_accession" in identifier
+                    ),
+                    None,
+                )
+            except TypeError:
+                continue
+            if not seq_id:
+                continue
+
+            try:
+                tax_id = next(
+                    (
+                        identifier["ncbi_taxonomy_id"]
+                        for identifier in seq_organism
+                        if "ncbi_taxonomy_id" in identifier
+                    ),
+                    None,
+                )
+            except TypeError:
+                tax_id = None
+
+            sequence = entry["entity_poly"]["pdbx_seq_one_letter_code"]
+
+            if tax_id:
+                organism = Organism(
+                    id=str(tax_id),
+                    taxonomy_id=int(tax_id),
+                )
+            else:
+                organism = None
+
+            prot_record = ProteinRecord(
+                id=seq_id,
+                sequence=sequence,
+                organism=organism,
+                structure_id=structure_id,
+            )
+
+            if entry["rcsb_polymer_entity_feature"]:
+                for feature in entry["rcsb_polymer_entity_feature"]:
+                    if feature["type"] == "Pfam":
+                        for region in feature["feature_positions"]:
+                            region = prot_record.add_to_regions(
+                                id=feature["feature_id"],
+                                name=feature["type"],
+                                start=region["beg_seq_id"],
+                                end=region["end_seq_id"],
+                            )
+                            region.add_object_term(Annotation.FAMILY.value)
+
+            try:
+                for instance in entry["polymer_entity_instances"]:
+                    polymer_instances = instance.get(
+                        "rcsb_polymer_instance_feature", None
+                    )
+                    if not polymer_instances:
+                        continue
+                    for feature in instance["rcsb_polymer_instance_feature"]:
+                        if feature["name"] == "sheet":
+                            for strand in feature["feature_positions"]:
+                                region = prot_record.add_to_regions(
+                                    name=feature["name"],
+                                    start=strand["beg_seq_id"],
+                                    end=strand["end_seq_id"],
+                                )
+                                region.add_object_term(Annotation.BETASTRAND.value)
+                                region.add_object_term(
+                                    "http://edamontology.org/topic_3542"
+                                )  # secondary structure
+
+                        elif feature["name"] == "helix":
+                            for helix in feature["feature_positions"]:
+                                region = prot_record.add_to_regions(
+                                    name=feature["name"],
+                                    start=helix["beg_seq_id"],
+                                    end=helix["end_seq_id"],
+                                )
+                                region.add_object_term(Annotation.ALPHAHELIX.value)
+                                region.add_object_term(
+                                    "http://edamontology.org/topic_3542"
+                                )  # secondary structure
+
+                        elif feature["name"] == "binding_site":
+                            positions = [
+                                site["beg_seq_id"]
+                                for site in feature["feature_positions"]
+                            ]
+                            site = prot_record.add_to_sites(
+                                name=feature["name"], positions=positions
+                            )
+                            site.add_object_term(Annotation.BINDING_SITE.value)
+
+            except KeyError:
+                pass
+
+            records.append(prot_record)
+
+        return records