Merge pull request #183 from pbashyal-nmdp/shortnull_for_exon

Support short null for `exon` mode

Author: mmaiers-nmdp

pyard/data_repository.py

@@ -362,17 +362,12 @@ def generate_alleles_and_xx_codes_and_who(
         xx_codes = db.load_dict(db_connection, "xx_codes", ("allele_1d", "allele_list"))
         xx_codes = {k: v.split("/") for k, v in xx_codes.items()}
 
-        shortnulls = db.load_dict(
-            db_connection, "shortnulls", ("shortnull", "allele_list")
-        )
-        shortnulls = {k: v.split("/") for k, v in shortnulls.items()}
-
         exp_alleles = db.load_dict(
             db_connection, "exp_alleles", ("exp_allele", "allele_list")
         )
         exp_alleles = {k: v.split("/") for k, v in exp_alleles.items()}
 
-        return valid_alleles, who_alleles, xx_codes, who_group, shortnulls, exp_alleles
+        return valid_alleles, who_alleles, xx_codes, who_group, exp_alleles
 
     # Create a Pandas DataFrame from the mac_code list file
     # Skip the header (first 6 lines) and use only the Allele column
@@ -481,6 +476,17 @@ def generate_alleles_and_xx_codes_and_who(
         db_connection, "who_group", flat_who_group, columns=("who", "allele_list")
     )
 
+    return valid_alleles, who_alleles, xx_codes, who_group, exp_alleles
+
+
+def generate_short_nulls(db_connection, who_group):
+    if db.table_exists(db_connection, "shortnulls"):
+        shortnulls = db.load_dict(
+            db_connection, "shortnulls", ("shortnull", "allele_list")
+        )
+        shortnulls = {k: v.split("/") for k, v in shortnulls.items()}
+        return shortnulls
+
     # shortnulls
     # scan WHO alleles for those with expression characters and make shortnull mappings
     # DRB4*01:03N | DRB4*01:03:01:02N/DRB4*01:03:01:13N
@@ -508,8 +514,7 @@ def generate_alleles_and_xx_codes_and_who(
 
     db.save_dict(db_connection, "shortnulls", shortnulls, ("shortnull", "allele_list"))
     shortnulls = {k: v.split("/") for k, v in shortnulls.items()}
-
-    return valid_alleles, who_alleles, xx_codes, who_group, shortnulls, exp_alleles
+    return shortnulls
 
 
 def generate_mac_codes(db_connection: sqlite3.Connection, refresh_mac: bool):

pyard/pyard.py

@@ -30,7 +30,7 @@
 from . import data_repository as dr
 from .smart_sort import smart_sort_comparator
 from .exceptions import InvalidAlleleError, InvalidMACError, InvalidTypingError
-from .misc import get_n_field_allele, get_2field_allele
+from .misc import get_n_field_allele, get_2field_allele, expression_chars
 
 HLA_regex = re.compile("^HLA-")
 
@@ -109,12 +109,14 @@ def __init__(
             self.who_alleles,
             self.xx_codes,
             self.who_group,
-            self.shortnulls,
             self.exp_alleles,
         ) = dr.generate_alleles_and_xx_codes_and_who(
             self.db_connection, imgt_version, self.ars_mappings
         )
 
+        # Generate short nulls from WHO mapping
+        self.shortnulls = dr.generate_short_nulls(self.db_connection, self.who_group)
+
         # Load Serology mappings
         dr.generate_serology_mapping(self.db_connection, imgt_version)
         # Load V2 to V3 mappings
@@ -213,7 +215,16 @@ def redux(self, allele: str, redux_type: VALID_REDUCTION_TYPES, reping=True) ->
                 return allele
         elif redux_type == "exon":
             if allele in self.ars_mappings.exon_group:
-                return self.ars_mappings.exon_group[allele]
+                exon_group_allele = self.ars_mappings.exon_group[allele]
+                # Check if the 3 field exon allele has a 4 field alleles
+                # that all have the same expression characters
+                last_char = allele[-1]
+                if last_char in expression_chars:
+                    exon_short_null_allele = exon_group_allele + last_char
+                    if self.is_shortnull(exon_short_null_allele):
+                        return exon_short_null_allele
+
+                return exon_group_allele
             else:
                 # for 'exon' return allele with only first 3 fields
                 return ":".join(allele.split(":")[0:3])
@@ -360,7 +371,7 @@ def redux_gl(self, glstring: str, redux_type: VALID_REDUCTION_TYPES) -> str:
             else:
                 raise InvalidMACError(f"{glstring} is an invalid MAC.")
 
-        # Handle shortnulls
+        # Handle short nulls
         if self._config["reduce_shortnull"] and self.is_shortnull(glstring):
             return self.redux_gl("/".join(self.shortnulls[glstring]), redux_type)
 

tests/features/shortnulls.feature

@@ -1,20 +1,27 @@
-Feature: shortnull
+Feature: Short Nulls
+
+  **Short Nulls**: If a reduced allele with an expression character has the same expression
+  character in it's 4 field expansion, the expression character should be propagated in the
+  reduced version of the allele.
 
   Scenario Outline:
 
     Given the allele as <Allele>
     When reducing on the <Level> level (ambiguous)
     Then the reduced allele is found to be <Redux Allele>
 
+    Examples: expression characters not propagated
+      | Allele         | Level | Redux Allele                        |
+      | DRB4*01:03N    | lgx   | DRB4*01:01                          |
+      | DRB4*01:03:01N | lgx   | DRB4*01:01                          |
+      | DRB5*01:08N    | lgx   | DRB5*01:02/DRB5*01:08               |
 
-    Examples: shortnulls
-      | Allele         | Level | Redux Allele                |
-      | DRB4*01:03N    | lgx   | DRB4*01:01                  |
-      | DRB4*01:03N    | exon  | DRB4*01:03:01               |
+    Examples: expression characters propagated
+      | Allele         | Level | Redux Allele                        |
+      | DRB4*01:03N    | exon  | DRB4*01:03:01N                      |
       | DRB4*01:03N    | W     | DRB4*01:03:01:02N/DRB4*01:03:01:13N |
-      | DRB4*01:03:01N | lgx   | DRB4*01:01                  |
-      | DRB4*01:03:01N | exon  | DRB4*01:03:01               |
+      | DRB4*01:03:01N | exon  | DRB4*01:03:01N                      |
       | DRB4*01:03:01N | W     | DRB4*01:03:01:02N/DRB4*01:03:01:13N |
-      | DRB5*01:08N    | lgx   | DRB5*01:02/DRB5*01:08 |
-      | DRB5*01:08N    | exon  | DRB5*01:08:01N/DRB5*01:08:02N |
-      | DRB5*01:08N    | W     | DRB5*01:08:01N/DRB5*01:08:02N |
+      | DRB5*01:08N    | exon  | DRB5*01:08:01N/DRB5*01:08:02N       |
+      | DRB5*01:08N    | W     | DRB5*01:08:01N/DRB5*01:08:02N       |
+      | A*01:04N       | exon  | A*01:04:01N                         |