Batch Reductions Update:

- `homozygosify_glstring`: When creating a GL String if one of the typ allele(s) is not available, homozygosify from the other typ
- `suppress_reduced_locus_column`: When creating GL String, you may not want the reduced locus columns in the output.

Author: pbashyal-nmdp

extras/reduce_conf.json

@@ -123,6 +123,8 @@
   "new_column_for_redux": true,
   "reduced_column_prefix": "reduced_",
   "generate_glstring": true,
+  "homozygosify_glstring": true,
+  "suppress_reduced_locus_column": true,
   "output_file_format": "csv",
   "apply_compression": "gzip",
   "verbose_log": true

extras/reduce_conf_glstring.json

@@ -27,6 +27,8 @@
   "new_column_for_redux": true,
   "reduced_column_prefix": "reduced_",
   "generate_glstring": true,
+  "homozygosify_glstring": true,
+  "suppress_reduced_locus_column": true,
   "output_file_format": "csv",
   "apply_compression": "gzip",
   "verbose_log": true

extras/sample.csv

@@ -1,4 +1,4 @@
 rid,did,r_a_typ1,r_a_typ2,r_b_typ1,r_b_typ2,r_c_typ1,r_c_typ2,r_drb1_typ1,r_drb1_typ2,r_dpb1_typ1,r_dpb1_typ2,d_a_typ1,d_a_typ2,d_b_typ1,d_b_typ2,d_c_typ1,d_c_typ2,d_drb1_typ1,d_drb1_typ2,d_dpb1_typ1,d_dpb1_typ2,r_drb3_typ1,r_drb3_typ2,r_drb4_typ1,r_drb4_typ2,r_drb5_typ1,r_drb5_typ2,d_drb3_typ1,d_drb3_typ2,d_drb4_typ1,d_drb4_typ2,d_drb5_typ1,d_drb5_typ2
-2110,123,A*01:AB,A*29:79,B*18:67,B*51:275,C*05:01:19,C*02:85:02,DRB1*03:03,DRB1*14:144,DPB1*193:01:01,DPB1*582:01:01,A*01:AB,A*29:79,B*18:67,B*51:275,C*05:01:19,C*02:85:02,DRB1*03:03,DRB1*14:144,DPB1*193:01:01,DPB1*582:01:01,DRB3*02:189,DRB3*03:09,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,DRB5*01:93,DRB5*02:02:01
+2110,123,A*01:AB,,B*18:67,B*51:275,C*05:01:19,C*02:85:02,DRB1*03:03,DRB1*14:144,DPB1*193:01:01,DPB1*582:01:01,A*01:AB,A*29:79,B*18:67,B*51:275,C*05:01:19,C*02:85:02,DRB1*03:03,DRB1*14:144,DPB1*193:01:01,DPB1*582:01:01,DRB3*02:189,DRB3*03:09,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,DRB5*01:93,DRB5*02:02:01
 2111,456,A*01:01:42,A*30:12:02,B*44:02:32,B*35:42,C*03:148,C*04:322,DRB1*13:01:16,DRB1*15:80N,DPB1*914:01:01,DPB1*278:01:01,A*01:01:42,A*30:12:02,B*44:02:32,B*35:42,C*03:148,C*04:322,DRB1*13:01:16,DRB1*15:80N,DPB1*914:01:01,DPB1*278:01:01,NNNN,NNNN,DRB4*01:53,DRB4*01:31,NNNN,NNNN,NNNN,NNNN,NNNN,NNNN,DRB5*01:102,DRB5*01:103
 2113,789,A*02:247,A*03:227,B*15:570,B*07:02:01:17,C*16:01:10,C*06:102,DRB1*13:156,DRB1*14:167:01,DPB1*405:01:01,DPB1*479:01:01,A*02:247,A*03:227,B*15:570,B*07:02:01:17,C*16:01:10,C*06:102,DRB1*13:156,DRB1*14:167:01,DPB1*405:01:01,DPB1*479:01:01,NNNN,NNNN,DRB4*01:79,DRB4*01:119,NNNN,NNNN,DRB3*02:189,DRB3*03:09,NNNN,NNNN,NNNN,NNNN

extras/sample_glstring.csv

@@ -1,3 +1,3 @@
 rid,did,recip_gl,donor_gl
-123,456,A*02:GNF+A*03:XYZ^B*07:ABD+B*44:AWA,A*02:01:01+A*03:01:01^B*07:RVXR+B*44:XYAG
+123,456,A*02:GNF+A*03:AB^B*07:ABD+B*44:AWA,A*02:01:01+A*03:01:01^B*07:RVXR+B*44:XYAG
 789,345,A*01:TUS+A*24:02:01G^B*08:ARGR+B*08:ARGS,A*02:01:01+A*01:PXTD^B*51:01:01G+B*40:BWUP

scripts/pyard-reduce-csv

@@ -231,22 +231,44 @@ def reduce_locus_columns(df, ard_config, locus_column_mapping, verbose):
             for locus in locus_column_mapping[subject]:
                 slug_column = locus + "_slug"
                 slug_columns.append(slug_column)
-                if len(locus_column_mapping[subject][locus]) > 1:
-                    df[slug_column] = (
-                        df[locus_column_mapping[subject][locus][0]]
-                        + "+"
-                        + df[locus_column_mapping[subject][locus][1]]
+                locus_typ_pair = locus_column_mapping[subject][locus]
+                if len(locus_typ_pair) > 1:
+                    df[slug_column] = df[locus_typ_pair].apply(
+                        create_reduced_slug, axis=1
                     )
                 else:
                     df[slug_column] = df[locus_column_mapping[subject][locus][0]]
 
+                if ard_config.get("suppress_reduced_locus_column"):
+                    df.drop(columns=locus_typ_pair, inplace=True)
+
             df[subject + "_gl"] = df[slug_columns].agg("^".join, axis=1)
             df[subject + "_gl"] = df[subject + "_gl"].apply(
                 lambda gl: gl.replace("^+", "")
             )
             df.drop(columns=slug_columns, inplace=True)
 
 
+def create_reduced_slug(locus_typ1_typ2_pair):
+    typ1 = locus_typ1_typ2_pair.iloc[0]
+    typ2 = locus_typ1_typ2_pair.iloc[1]
+
+    if not typ1 and not typ2:
+        return ""
+
+    if typ1 and typ2:
+        return typ1 + "+" + typ2
+    elif ard_config.get("homozygosify_glstring"):
+        if typ1:
+            return typ1 + "+" + typ1
+        if typ2:
+            return typ2 + "+" + typ2
+    else:
+        if typ2:
+            return typ2
+    return typ1
+
+
 def reduce_glstring(glstring: str) -> str:
     try:
         return ard.redux(glstring, ard_config["redux_type"])