Merge pull request #238 from pbashyal-nmdp/serology_redux

Support Reduction to Serology

Author: mmaiers-nmdp

README.md

@@ -176,6 +176,7 @@ ard.redux('B14', 'lg')
 | `W`            | Reduce/Expand to 3 field WHO nomenclature level |
 | `exon`         | Reduce/Expand to exon level                     |
 | `U2`           | Reduce to 2 field unambiguous level             |
+| `S`            | Reduce to Serological level                     |
 
 ### Perform DRB1 blending with DRB3, DRB4 and DRB5
 

pyard/ard.py

@@ -33,6 +33,7 @@
 from .misc import (
     get_n_field_allele,
     get_2field_allele,
+    is_2_field_allele,
     validate_reduction_type,
 )
 from .constants import (
@@ -253,6 +254,24 @@ def _redux_allele(
             else:
                 # If ambiguous, reduce to G group level
                 return self._redux_allele(allele, "lgx")
+        elif redux_type == "S":
+            # find serology equivalent in serology_mapping
+            serology_mapping = db.find_serology_for_allele(self.db_connection, allele)
+            serology_set = set()
+            if is_2_field_allele(allele):
+                for serology, allele_list in serology_mapping.items():
+                    allele_list_lgx = self.redux(allele_list, "lgx")
+                    if allele in allele_list_lgx:
+                        serology_set.add(serology)
+            else:
+                for serology, allele_list in serology_mapping.items():
+                    if allele in allele_list:
+                        serology_set.add(serology)
+            return "/".join(
+                sorted(
+                    serology_set, key=functools.cmp_to_key(self.smart_sort_comparator)
+                )
+            )
         else:
             # Make this an explicit lookup to the g_group or p_group table
             # for stringent validation
@@ -293,7 +312,7 @@ def _sorted_unique_gl(self, gls: List[str], delim: str) -> str:
         all_gls = []
         for gl in gls:
             all_gls += gl.split(delim)
-        unique_gls = set(all_gls)
+        unique_gls = filter(lambda s: s != "", set(all_gls))
         return delim.join(
             sorted(unique_gls, key=functools.cmp_to_key(self.smart_sort_comparator))
         )

pyard/constants.py

@@ -25,7 +25,7 @@
 
 HLA_regex = re.compile("^HLA-")
 
-VALID_REDUCTION_TYPES = ["G", "P", "lg", "lgx", "W", "exon", "U2"]
+VALID_REDUCTION_TYPES = ["G", "P", "lg", "lgx", "W", "exon", "U2", "S"]
 expression_chars = ["N", "Q", "L", "S"]
 # List of P and G characters
 PandG_chars = ["P", "G"]

pyard/db.py

@@ -370,6 +370,27 @@ def similar_alleles(connection: sqlite3.Connection, allele_name: str) -> Set[str
     return alleles
 
 
+def find_serology_for_allele(
+    connection: sqlite3.Connection, allele_name: str
+) -> Dict[str, str]:
+    """
+    Find similar alleles starting with the provided allele_name.
+
+    :param connection: db connection of type sqlite.Connection
+    :param allele_name: Allele name to use as a prefix to find similar alleles
+    :return: list of similar alleles
+    """
+    query = (
+        "SELECT serology, allele_list FROM serology_mapping WHERE allele_list LIKE ?"
+    )
+    cursor = connection.execute(query, (f"%{allele_name}%",))
+    results = cursor.fetchall()
+    # fetchall() returns a list of tuples of results
+    # e.g. [('A1', ''A*01:01:01:01/A*01:01:01:03')]
+    serology_mapping = {serology: allele_list for serology, allele_list in results}
+    return serology_mapping
+
+
 def get_user_version(connection: sqlite3.Connection) -> int:
     """
     Retrieve user_version from db

pyard/misc.py

@@ -67,6 +67,10 @@ def number_of_fields(allele: str) -> int:
     return len(allele.split(":"))
 
 
+def is_2_field_allele(allele: str) -> bool:
+    return number_of_fields(allele) == 2
+
+
 # computes a valid G name based on the ambiguity string
 def get_G_name(a: str) -> str:
     a = a.split("/")[0]

pyard/smart_sort.py

@@ -61,7 +61,11 @@ def smart_sort_comparator(a1, a2):
     if a1 == a2:
         return 0
 
-    # Extract and Compare first fields first
+    # Handle serology
+    if ":" not in a1:
+        return 1 if a1 > a2 else -1
+
+    # Extract and Compare 1st fields first
     a1_f1 = int(a1[a1.find("*") + 1 : a1.find(":")])
     a2_f1 = int(a2[a2.find("*") + 1 : a2.find(":")])
 

requirements.txt

@@ -1,2 +1,3 @@
 toml==0.10.2
+numpy==1.24.3
 pandas==1.5.3

tests/features/serology_redux.feature

@@ -0,0 +1,21 @@
+Feature: Serology Reduction
+
+  For a GL String, find the serological equivalents
+  Serological reduction mode is *S*
+
+  Scenario Outline: Serology Reduction
+
+    Given the allele as <Allele>
+    When reducing on the <Level> level with ping
+    Then the reduced allele is found to be <Redux Allele>
+
+    Examples:
+      | Allele                          | Level | Redux Allele        |
+      | A*01:01:01:01                   | S     | A1                  |
+      | A*01:01                         | S     | A1                  |
+      | A*01:AABJE                      | S     | A1/A36              |
+      | A*03:XX                         | S     | A3                  |
+      | B*44:02:01:11/B*44:02:01:12     | S     | B12/B44             |
+      | B*13:03                         | S     | B13                 |
+      | B*13:04                         | S     | B15/B21             |
+      | B*15:01/B*15:02/B*15:03/B*15:04 | S     | B15/B62/B70/B72/B75 |